Testing the Local Enumerator Approach for Agricultural Data Collection By Christopher Root A THESIS Submitted to Michigan State University in partial fulfilment of the requirements for a degree of Agricultural, Food and Resource Economics – Masters of Science 2017 ABSTRACT Testing the Local Enumerator Approach for Agricultural Data Collection By Christopher Root Technology adoption is widely regarded as critical to agricultural development. However, survey based data on technology adoption is costly to collect. This research is an attempt to lower this cost by using locally based enumerators and tablets. The hypothesis this research tests is that the ‘local enumerator approach’ will reduce costs while maintaining data quality. This hypothesis is tested by comparing adoption data collected through the local enumerator approach with that collected through a conventional survey in India. Means comparison tests indicate statistically significant differences in adoption estimates derived from the two approaches. Regression analysis, controlling for village fixed effects, covariates and enumerator fixed effects, is used to identify adoption measurement differences between the local and the conventional enumerator approaches. However, none of these analytical approaches are able to eliminate all the differences in adoption estimates, implying significant differences in data quality generated by these two approaches. The study design however was not able to control several potential confounding factors, such as enumerator training method, differences in questionnaire design, and data collection application tools that may have influenced data quality. Although costs are found to be comparable between the two approaches, over the long run, there is potential for costs of local enumerator approach to decrease relative to the conventional approach because of cost efficiencies. However, more effort is needed to ensure data quality before this approach can be considered a cost-effective and a reliable method of data collection in developing countries. ACKNOWLEDGEMENTS I wish to gratefully acknowledge the patient support of my professor Dr. Mywish Maredia, as well as my committee members Dr. Andrew Dillon and Dr. Susan Wyche. Additionally, I appreciate the helpfulness of the survey implementing firms. Finally, I wish to acknowledge the Standing Panel on Impact Assessment (SPIA) of the Consultative Group on International Agricultural Research’s (CGIAR) Independent Science and Partnership Council (ISPC) for funding this research through the Strengthening Impact Assessment in the CGIAR (SIAC) project. iii TABLE OF CONTENTS LIST OF TABLES ......................................................................................................................... vi 1. INTRODUCTION ................................................................................................................... 1 1.1 Motivation ............................................................................................................................. 1 1.2. The Local Enumerator Approach ......................................................................................... 2 Advantages .............................................................................................................................. 3 Disadvantages .......................................................................................................................... 4 Sampling differences ............................................................................................................... 5 Overview of research and objectives ....................................................................................... 5 Layout of paper ........................................................................................................................ 7 2. LITERATURE REVIEW ........................................................................................................ 8 2.1 Local enumerators ................................................................................................................. 8 2.2 Mobile phone or tablet based data collection ........................................................................ 9 2.3 Measurement error .............................................................................................................. 11 Sources of measurement error ............................................................................................... 11 Measurement error and the regression model........................................................................ 13 Measurement error and development research ...................................................................... 15 Statistical and practical significance...................................................................................... 17 3. APPROACHES EVALUATED ............................................................................................ 19 3.1 Agricultural technologies studied........................................................................................ 20 3.2 Survey implementers ........................................................................................................... 22 LEA Implementer 1 ............................................................................................................... 24 LEA Implementer 2 ............................................................................................................... 24 LEA Implementer 3 ............................................................................................................... 25 Comparison ............................................................................................................................ 26 CAPI tools ............................................................................................................................. 26 4. FRAMEWORK, METHODOLOGY AND DATA .............................................................. 28 4.1 Framework and Methodology ............................................................................................. 28 iv 4.2 Data ..................................................................................................................................... 31 4.3 Sample size.......................................................................................................................... 33 4.4 Sampling method................................................................................................................. 34 4.5 Costs .................................................................................................................................... 37 5. RESULTS .............................................................................................................................. 38 5.1 Comparison of descriptive statistics.................................................................................... 38 Household characteristics and adoption estimates ................................................................ 38 Comparison of enumerators’ characteristics ......................................................................... 42 5.2 Balancing tests for covariates .............................................................................................. 43 5.3 Identifying adoption differences ......................................................................................... 45 5.4 Comparison of missing values ............................................................................................ 49 5.5 Exploring the factors contributing to differences in data quality ........................................ 50 5.6 Cost...................................................................................................................................... 53 6. DISCUSSION AND CONCLUSION ................................................................................... 55 6.1 Summary of findings ........................................................................................................... 55 6.2 Sources of differences in adoption estimates ...................................................................... 57 6.3. Feasibility of the local enumerator approach ..................................................................... 59 REFERENCES ............................................................................................................................. 61 APPENDICES .............................................................................................................................. 65 APPENDIX A: LEA Implementer 1 Questionnaire ................................................................. 66 APPENDIX B: CEA Implementer – Groundnuts Questionnaire.............................................. 74 APPENDIX C: LEA Implementer 2 Questionnaire .................................................................. 89 APPENDIX D: LEA Implementer 3 Questionnaire ............................................................... 105 APPENDIX E: CEA Implementer – Wheat and Rice Questionnaire ..................................... 117 v LIST OF TABLES Table 1: Implementers, districts and cropping system studied ..................................................... 20 Table 2: Summary of differences between survey implementation models ................................ 23 Table 3: Technologies and adoption outcome variables included in the comparative analysis by implementers ................................................................................................................................. 32 Table 4: List of variables common across the four implementer datasets .................................... 33 Table 5: Sample size by implementer and district ........................................................................ 34 Table 6: Comparison of household characteristics between the three pairs of LEA and CEA approaches..................................................................................................................................... 39 Table 7: Adoption rate and adoption area for different types of technologies: Comparison between the two pairs of survey approaches by implementers ..................................................... 41 Table 8: Comparison of enumerator characteristics ..................................................................... 43 Table 9: Balancing tests for covariates with village fixed effects ................................................ 45 Table 10: Differences in the mean estimates of adoption variables derived from LEA and CEA datasets: Comparison of results across three methods .................................................................. 47 Table 11: Average number of missing values (out of ten covariates) per respondent .................. 49 Table 12: Missing value differences and coefficient comparison ................................................ 50 Table 13: Enumerator characteristics coefficients with total technology adoption area dependent variable .......................................................................................................................................... 52 Table 14: Enumerator characteristics coefficients with missing values dependent variable ........ 52 Table 15: Actual costs for survey implementation (USD)............................................................ 53 vi ABBREVIATIONS CAPI Computer Assisted Personal Interviews CEA Conventional Enumerator Approach CIMMYT International Maize and Wheat Improvement Center DSR Direct seeded rice GoI Government of India ICRISAT The International Crops Research Center for Semi-Arid Tropics IECC Intra-enumerator correlation coefficient LEA Local Enumerator Approach SRS Simple Random Sampling ZT Zero till vii 1. INTRODUCTION 1.1 Motivation Despite increasing urbanization, agriculture still has an important role to play in poverty alleviation and economic development. Seventy percent of the world’s poor live in rural areas and are mostly dependent on agriculture for their livelihoods (World Bank 2015a). Likewise 70% of the population in countries classified by the World Bank as Low Income live in rural areas, indicating the limited extent of structural transformation in these countries (World Bank 2015b). Increasing agricultural productivity is key to both poverty alleviation and economic development. Improved agricultural technology is critically important for increasing agricultural productivity. Transferring improved technologies to farmers in developing countries has long been a priority of governments and international organizations alike. However, farmers may not adopt a technology targeted to them for a variety of reasons which are well-documented in the literature (see for example Jack 2013; Feder & Umali 1993). To accelerate dissemination of improved technologies, governments and organizations need to track adoption and better understand reasons for farmers’ adoption decisions. However data on adoption of agricultural technologies in developing countries is limited and often out of date. Adoption studies are often based on micro surveys which are not nationally representative and cross sectional data which does not show the dynamics of adoption over time (Doss, 2006 & Feder & Umali, 1993). Attempts to measure adoption on larger scale and with lower cost have often been based on aggregation of expert estimations of adoption for the 1 location and crops within their focus geographies (Walker and Alwang 2015). For obvious reasons this approach is likely to lead to imprecise measurement. Data collected directly from a representative sample of farmers is considered to be more accurate than data expert solicitation. Some type of farmer-level data requires more frequent collection to reduce recall bias, such as seasonal data related to area, production or transactional data including prices. Additionally, for tracking and monitoring, adoption and other data needs to be collected on a regular basis to be able to be used for timely policy decision-making. The primary constraint to collecting data from the farmers, and doing so more frequently (or on a regular basis), is the cost of collecting that data. A reduction in the cost of collecting data implies a proportionate increase in the amount of data that can be collected within a fixed budget. This research tests a model that aims to reduce data collection costs - at least in the longrun - thereby potentially increasing the frequency of data collection. 1.2. The Local Enumerator Approach The integration of communication and digital technologies such as mobile phones and tablets in conventional surveys can potentially reduce survey costs, speed up collection time and improve data quality (Caeyers et al., 2012). However even with these technologies there are costs that could potentially be eliminated. This is because the conventional enumerator approach usually depends on enumerators hired from urban areas who, after training, are sent individually or in small teams from village to village in a sampling area. Transportation costs, labor and per diem 2 therefore must be paid both to get enumerators to and from survey areas as well as between enumeration clusters within survey areas. The approach tested through this research is to utilize local enumerators who live or are based in or near agricultural communities to complete interviews of farmers in their communities. With the local enumerator approach, data is collected by local enumerators using a tablet or smartphone and uploaded to a database through the internet. Once trained, these local enumerators could be available to carry out a range of agricultural data collection assignments for different data users. The local enumerator service could be managed by a domestic firm or organization and depending on the demand, could expand over time. This research tests three different practical applications of this envisioned local enumerator approach in India. Advantages The proposed research attempts to verify the effectiveness of the local enumerator approach (LEA) in measuring agricultural adoption. To be effective, the approach should yield comparable results to a conventional enumerator approach (CEA) at lower costs. The lower cost will likely be a result of eliminating enumerator transport and per diem costs. Additionally, implementation speed could be faster due to reduced time needed for reoccurring training and transport. In addition to potential cost savings there are potential methodological advantages associated with the LEA. For one, the LEA makes it easier to carry out multiple rounds of data collection. McKenzie (2012) shows that when autocorrelation is low, power can be gained by carrying out multiple post-treatment measurements. Chao et al. (2012) find that revisits - that is returning to the business or household if the primary respondent is not available – have significant effects in 3 reducing bias and mean squared error of sample estimates. Revisiting respondents will obviously be easier for local enumerators than enumerators typically hired under the conventional approach. Finally, there may be some questions for which local enumerators can solicit more accurate responses based on their trustworthiness or local knowledge. An example is a local agricultural practice that outsiders may have difficulty explaining. Disadvantages There are also potential disadvantages that need to be taken into consideration. In general, they imply potentially lower quality data resulting from the local enumerator approach. These relate to the quality of enumerators the LEA is able to attract, lack of supervision during data collection, and the sampling method used to identify the respondents. It is not certain that the caliber of enumerators available in rural communities is as high as those typically available in cities from where survey firms typically draw enumerators in the conventional approach. Additionally, the LEA is more decentralized than the CEA and therefore quality assurance may be more challenging. Also, in the LEA the last tiers of sample selection (e.g., villages and farmers within the villages) need to be made by the enumerator, and it is possible that the sampling may be based more on convenience than on the principles of randomness. This research attempts to address these potential disadvantages and how it affects the quality of the data collected. Another potential disadvantage of LEA is that for some types of questions respondents are less likely to be forthright with a community member (if the enumerator ‘local’) than with a stranger (as is likely with the conventional approach). An example is agricultural income. Finally, there is the issue of enumerator attrition having a stronger effect on the sample then in the case of a 4 mobile enumerator based survey. This is because with a conventional survey, if an enumerator drops out, other enumerators can be used to cover the households of the missing enumerator. Local enumerators however are responsible for a fixed local area and may have other work or domestic responsibilities that prohibit them from travelling elsewhere. Sampling differences The theoretical difference between statistical precision or power between the CEA and the LEA is unclear. This research utilized a two-stage cluster design with villages as clusters. In that case the difference in statistical power is a function of the number of clusters with more clusters leading to more statistical power. For the CEA, the number of clusters given a fixed budget is a function of adding an additional cluster which includes transport, additional wages and lodging. For the LEA, the number of clusters is likely a function of the cost of training additional LEs as local enumerators are by definition limited in the radius in which they can enumerate. Therefore, sampling differences depend on differences in cost structures between the two approaches that incentivize different survey design decisions. Overview of research and objectives This research attempts to verify the effectiveness of the LEA in measuring agricultural adoption. As mentioned, to be effective the LEA should yield comparable results to a CEA at lower costs. Therefore, this research tests whether or not indicators measured through the LEA are statistically indistinguishable from those measured through the CEA and compares the costs of the two survey methods. It also looks at whether measured differences are practically different in magnitude. Then it looks for differences in enumerators’ characteristics across the LEA and CEA enumerators to see if that may explain measurement differences. 5 Three variations of the local enumerator approach are evaluated. These variations are based on the interpretations of the model by three different implementing partners. The three partners were awarded funding to implement their version of the local enumerator approach based on their responses to a solicitation specifying basic parameters of our vision of the local enumerator approach. Therefore this research does not strictly evaluate the LEA as conceptualized and described above. Instead, by way of comparison to a conventional survey, this research assesses the effectiveness of the conception and implementation of the three different versions of the LEA in India. The three different implementers of the LEA conduct studies of adoption of technologies in five different districts in India. Each implementer surveyed between 600 and 800 farmers per district in two districts. In Andhra Pradesh in Southern India, one implementer conducted surveys of water-conserving groundnut technologies in two districts. In Haryana and Punjab in Northern India, another implementer conducted surveys of wheat farming system technologies. Finally, in Haryana and Bihar in Northern India, a third LEA implementer conducted two surveys on also on wheat farming technologies. The CEA implementer conducted five surveys on adoption of the same populations in the same districts using similar sample sizes. The analysis compares adoption measurement for each LEA with results measured by the CEA for the same population. After identifying statistically significant differences in means for adoption and adoption covariates between the different samples, village fixed effects and adoption covariates are used to account for unobservable village level differences that may affect 6 adoption estimate differences. Following this, enumerator fixed effects along with adoption covariates are used to attempt to explain the remaining differences in adoption estimates. Additional steps taken include a comparison of missing values across implementers as well as regressing enumerator characteristics on adoption outcomes and missing values to identify enumerator characteristics potentially responsible for differences between groups. Finally, cost comparisons are made based on actual survey costs for the four survey implementers. Layout of paper This introduction is followed by a relatively brief literature review that covers literature on three related topics: the local enumerator approach, computer assisted personal interviews (CAPI) and measurement error. The discussion of the literature on measurement error is subdivided into three parts: sources of measurement error, regression implications of measurement error and examples of measurement error related development research. Chapter three contains a description of the four implementers (3 LEA and 1 CEA) and the approaches they used to their surveys including sampling. It also describes the technologies whose adoption is evaluated here. Chapter 4 lays out the research framework, overviews methodologies, and provides a description of data and sampling to evaluate the local enumerator approach. Chapter 5 presents the results of the analysis. The final chapter discusses these results and draws conclusions. 7 2. LITERATURE REVIEW 2.1 Local enumerators Using enumerators who live within the same area as respondents is not new. National census surveys, including in India and the US, have been carried out by local or community enumerators (Smith et al. 2007 & Royce 1986). In India, these local enumerators are often teachers who conduct interviews outside of their regular work time (GOI 2011a). In the 2011 census, enumerators had three weeks to interview between 125 and 150 households. Supervisors were responsible for five to six enumerators (GOI 2011b). Researchers have also used local enumerators. For example, Udry’s research (1991) on credit in Northern Nigeria involved enumerators who lived in study villages over the one year study period and conducted monthly interviews under supervision from the author. Sitati et al. (2005) trained rural community members to act as enumerators for locals reporting crop raids by elephants. A health study in Tanzania utilized past project data collected from local teachers and health workers (Dotchin et al. 2008). In addition to censuses and research designs, development projects have utilized local enumerators. The Community Managed Sustainable Agriculture Project, a government run project in Andhra Pradesh, India, reaches one million farmers in 11,000 villages. The program provides “community professionals” with information on sustainable agricultural practices to disseminate in their communities. Community professionals also collect data via cell phone on the effectiveness of the programs they are implementing (World Bank 2013). The Community Knowledge Worker (CKW) program, implemented in Uganda by the Grameen Foundation, trains CKWs to assist farmers in their community through information they receive via their 8 mobile phones. To make the program fiscally sustainable when program funding ends, CKWs are trained to offer data collection services for other development organizations. However, demand for this service is not yet clear (USAID 2011; private discussion 2014). Finally, FAO and an NGO have set up a mobile-based drought early warning system in northern Uganda. To provide timely data, the system depends on monthly data sent via mobile phone that is collected by parish chiefs (WB 2013). A key difference between these three examples and the local enumerator approach proposed here is that the local workers in the three examples above are not paid exclusively to collect data. While not always explicit, the motivations for using local enumerators vary across these cases. For national censuses, using local enumerators is more logistically feasible than sending a team of enumerators around the country. In northern Nigeria, Udry (1991) was concerned about systematic non-sampling error introduced by the event-like nature of an interview with an outsider. In the case of the Sitati et al (2005) research, the use of local enumerators was imperative to be able to record observations as shortly as possible after they occurred, presumably resulting in more accurate data. Likewise, the drought early warning system depends on the timely data provided by local enumerators. Local enumerators also are advantageous for panel data collection involving multiple enumerator visits. They also allow for easier revisits if the targeted respondent is not initially available. 2.2 Mobile phone or tablet based data collection Caeyers et al. (2012) compare results from interviews conducted by pen and paper (also referred as PAPI—pen and paper based interviews) with computer assisted personal interviews (CAPI), 9 particularly for measuring household consumption. They look at both data errors as well as sample size reduction resulting from irreconcilable missing values. They do this by randomly assigning households to be interviewed either through pen and paper or through smartphone. The authors compare the number of data entry errors between the two methods and find that errors resulting in missing values are nearly eliminated through the electronic survey. A significant portion of this is because of the validation checks that can be built into CAPI. They find that CAPI reduced both the mean and variance of consumption estimates. They also find a negative effect of enumerator experience and education on survey errors and that this effect is smaller for the smartphone or tablet based survey. The implication of this finding is that for smartphone or tablet based surveys, the importance of the experience of the enumerator is reduced. This intuitive finding bodes well for the LEA. Nevertheless, the importance of training should not be underestimated. For mobile phone and tablet data collection, training costs may need to be even higher given that the technology itself must be learned (E-agriculture, 2012). Training costs are also higher where more complex data collection applications and phone technologies are used (WB, 2013). Several other lessons can be drawn from reviewing the experiences of others who have used mobile phone based data collection in developing countries. One is that the claimed advantages of cost savings and data integrity appear to be real. Cost savings come through reduced data cleaning cost because of better data integrity (USAID 2012). Data integrity can be improved through in digital questionnaires with skip logic as well as by the ability to reduce data fabrication through time stamps, geo-tagging and pictures (E-agriculture 2012). While these 10 technical approaches to reduce fabrication can be effective, it is also important to compensate enumerators well including giving bonuses for complete and accurate data submission (WB 2013). 2.3 Measurement error There is a growing body of literature on measurement issues in agricultural and other surveys in developing countries. Increasingly, researchers are aware that methodological choices in survey design have significant effects on research results. Researchers are detecting and quantifying these methodological measurement-related issues and their implications through randomized experiments similar to the one we propose here (McKenzie and Rosenzweig 2012). This chapter briefly reviews several of the most relevant ones to our research topic. However, before proceeding to describe these studies, the chapter discusses the sources of measurement error and their implications for regression analysis. Note that this research focuses primarily on measurement error or non-sampling errors. Sampling errors are any difference between the population N and the sample n. To correct for sampling error in complex surveys where simple random sampling has not been used such as cluster sampling, survey weights are needed and indeed are employed throughout this research (Brogan 2005). Sources of measurement error Three sources of non-random measurement error are relevant to this research: the questionnaire, the interviewer and the interviewee (Kasprzyk 2006 & SPO 2001). Note that data entry errors are expected to be minimized through the use of tablet based data collection (Caeyers et al. 2012). 11 The questionnaire can lead to several types of errors. This source of error is relevant to this research as each implementer used similar but not exactly the same questionnaire. One type of problem is specification problems, where concepts are not well defined and therefore ambiguous. A closely related class of problems is wording problems, where confusing words may cause misunderstanding on the part of the enumerator and the respondent. The length of the questionnaire can also cause measurement error as respondents lose focus resulting in less accurate responses over time. Similarly, question order can affect measurement error. Not only does this affect the time from the start of the questionnaire to the question, it also influences how much contextual information precedes a question. Question order also affects assimilation, which is the likelihood that a respondent’s answer to a question is similar to a response to a previous similar question. Response options can also introduce bias as categories chosen by questionnaire designers may omit important responses or the order of the categories influence which option respondents select (Kasprzyk 2006). Interviewers play a key role in data quality. This may be through their skill in following accurately the logic and intent of the questionnaire as well as their ability to elicit and record accurate information from the respondent by both making them comfortable and not influencing their response. However, little is known about the characteristics associated with a good enumerator. To some extent, enumerator effects can be reduced through training and supervision as well as questionnaire design and used of CAPI, as discussed above (Kasprzyk 2006). 12 The intra-enumerator correlation coefficient (IECC) is a measure that can be used to assess enumerator effects. This is defined as the ratio of the variance of an enumerator to that of all enumerators for the same variable. This ratio is typically around 0.02. An intuitive but important implication of the IECC is that fewer enumerators per survey increases the effect of an individual enumerator’s IECC on the precision of the entire sample. The reduction in precision is nonlinear, increasing more rapidly as enumerator interview counts increase (Kasprzyk, 2006). This implies a possible theoretical advantage for the local enumerator approach where enumerators are responsible for fewer clusters and potentially fewer respondents. Respondents also affect measurement error. There are two ways that are especially relevant to this research. Longer recall periods may introduce error as memory fades over time affecting the accuracy of, for example, responses to questions about a previous year. Another respondent source of measurement error is if the respondent is not the person in the household who knows the most about the topic they are being asked about (SPO 2001). For example, if respondent is not the head of the household, they may be less informed on the details of farm management. This research controls for this potential contributor to measurement error. Measurement error and the regression model Measurement error in the right-hand side variable can be represented as follows for the ordinary least squared (OLS) model satisfying the standard assumptions. 𝑦 = 𝛽𝑥 ∗ + 𝜇 (1) Where 𝑥 ∗ represents 𝑥 measured with error 𝜀 13 𝑥 = 𝑥∗ + 𝜀 (2) Where 𝜀 is not correlated with 𝑥 or 𝜇. Hausman (2001) shows that this measurement error in the right-hand side variable 𝑥 usually results in a downward bias of the estimator 𝛽. The extent of the bias then depends on the ratio of the variance of the measurement error 𝜀 to the variance of the true indicator 𝑥. When this ratio is higher, the bias of estimator 𝛽 is higher. Hausman (2001) cites empirical research showing downward bias most commonly in the range of 25 to 33 percent. Hausman (2001) also shows that measurement error in the left-hand side variable 𝑦 of an OLS regression does not introduce bias as the measurement error is accounted for by the error term 𝜇. However, the variance of the error term 𝜇 does increase, resulting in larger standard errors. The implication is that a larger sample is needed to get the same precision when there is measurement error in the left-hand side variable. The effect of mismeasurement of both right and left hand variables is generally the same as their individual effects discussed above so long as the two errors are uncorrelated with each other (Hausman 2001). In the case of binary left hand side variables for probit and logit models, measurement error or rather misclassification of binary variables can lead to biased and inconsistent estimators. However, maximum likelihood estimation provides consistent estimators so long as the combined probability of recording a 1 when the true value should be 0 plus the probability of recording a 0 when the true value should be 1 is less than 1 (Hausman 2001). 14 Measurement error and development research The design proposed here is similar to that used by Caeyers et al. (2012) to compare results from interviews conducted by pen and paper with computer assisted personal interviewing (CAPI). They do this by running their study alongside an already-planned conventional survey and randomly selecting additional households from the same enumerator areas. To isolate the effects of consistency checks – which is believed to be a critical feature of CAPI – the researchers randomly assigned CAPI respondents to two groups: a) full CAPI that includes consistency checks; and b) restricted CAPI that excludes consistency checks. The same enumerator team within an enumerator cluster conducted all three types of interviews. However, the interviews were scheduled to reduce time and interviewer clustering. The authors first look at data errors that result in missing values and consequently sample size reduction. These data errors include routing errors and unlikely or impossible data entries. They regress enumerator characteristics – years of schooling and experience with both PAPI and CAPI – on number of data errors. As discussed above, the authors find that the effects of enumerator education and experience are significant with PAPI but with CAPI these effects disappear. The implication is that enumerator experience and competency are less relevant with CAPI because of the internal consistency checks and skips/routing features (Caeyers et al. 2012). The authors used several models in their analysis. One was to regress a count variable defined as the number of problematic variables on dummy variables indicating assignment to the three groups. They expand this model to include household characteristics that the authors hypothesized would influence measurement error. Finally, the authors run the same model but replace household characteristics with enumerator characteristics. They confirm that their 15 analysis was robust to characteristics of the respondent, the interview, the interviewer and the location – the latter two through interviewer and cluster fixed effects (Caeyers et al. 2012). The authors look for classical measurement error in the explanatory variable (consumption) by testing for attenuation bias. To detect attenuation bias, they regress the log of consumption, the log of consumption interacted with a CAPI dummy, and a CAPI dummy alone on several schooling related dependent variables. They find the coefficient for consumption alone is smaller and statistically insignificant compared to consumption interacted with CAPI. The authors take this as evidence of attenuation bias and therefore classical measurement error in the PAPI version of the survey (Caeyers et al. 2012). Carletto et al. (2011) look to determine the impact of measurement error on the inverse land size hypothesis (the hypothesis is that smaller farms are more productive than larger ones). The authors use national data from Uganda, including land measurements done by GPS and by farmer estimate. They first provide means comparisons at the plot and household levels. They then use econometric analysis to examine a) the predictors of discrepancy in measurement and b) what impact that discrepancy has on estimating the inverse land size hypothesis. The authors then test the former by regressing the difference between the two estimates on a vector of variables they suspected influenced the farmers’ ability to estimate their land size including topography. For b) they run regressions using both land measurements and compare results using a Wald test. One of their key findings is that significant bias results from respondent rounding (approximating responses to a round number). They also find that the bias is inversely 16 proportionate to the size of the land holding. It is not clear if this can be generalized to mean that survey measurement error is more serious with smaller estimates. In a subsequent study, Carletto et al. (2013) expand the analysis to four African countries. Through means tests they find that farmer estimates are over-estimated for small land parcels and under estimated for large land holdings. Their econometric specification to estimate the predictors of bias differs from their previous study in that they add an indicator for GPSmeasured land area to control for bias of land area measurement. They also further expand the model to include dummies to represent rounding. Statistical and practical significance Statistical hypothesis testing such as student t tests (used here) predominates in sciences and social sciences. However, researchers caution against the overreliance on such statistical tests and perhaps more importantly, the conflation of statistical and practical significance. Statistical significance tests indicate whether or not a difference exists but say nothing about its magnitude. For both policy and science, whether or not a difference or causal relationship exists is less important than its magnitude (Ziliak and McCloskey 2008). For example, the difference between an agricultural technology adoption rates of 0.17 and 0.24 may be statistically significant but unlikely to be viewed as practically significant. That is to say, those two different adoption estimates should not elicit different policy responses. In such cases, it is not obvious how much more cost more precision is worth. While using local enumerators is not new, the idea proposed here to have a permanent infrastructure of professional local enumerators appears to be novel. However, measurement 17 error resulting in practical and not only statistically significant differences may undermine this approach. This may originate from the questionnaire, the numerator or the interviewee and can introduce bias into regression estimations. Combining the local enumerator approach with tabletbased data collection can be expected to reduce the effect of enumerator based measurement error through logic checks and validations. This should thereby mitigate potential differences in enumerator quality between local and conventional enumerators and thereby reduce data quality differences between the two approaches. 18 3. APPROACHES EVALUATED This section details the three enumeration approaches motivated based on the concept of the LEA approach, and one conventional approach, which is used as a comparison group. The conventional approach, as defined here, is that enumerators are hired from outside the survey area and sent around the survey area to conduct interviews during the survey period without returning home. By contrast, local enumerators are recruited from within or nearby the survey area and are able to return home every day. The three variations of the local enumerator approach were developed independently by three implementers in response to a solicitation issued by researchers at Michigan State University (MSU) laying out the basic parameters of the local enumerator approach. The fourth model, which is used as the comparison group was developed in response to a solicitation by the researchers at MSU for a conventional survey. Table 1 below shows the geographic scope of the surveys conducted and the focus technologies by each survey firm. The surveys focused on five 2districts in India—two southern districts (Anantapur and Kurnool) where groundnut is an important crop, and three northern districts (Karnal, Ludhiana and Vaishali) with wheat-based cropping system. In this chapter we describe the technologies included in this study for the wheat based and groundnut based farming systems. We then describe the four implementers selected for this study, their approaches, and sampling strategies. 19 Table 1: Implementers, districts and cropping system studied District Anantapur Kurnool Karnal Ludhiana Vaishali Region (State) LEA 1 LEA 2 LEA 3 Southern (Andhra Pradesh) Groundnuts Southern (Andhra Pradesh) Groundnuts Northern (Haryana) Wheat based Wheat based Northern (Punjab) Wheat based Northern (Bihar) Wheat based Comparison Groundnuts Groundnuts Wheat based Wheat based Wheat based 3.1 Agricultural technologies studied Wheat based farming system: Technologies promoted through Climate-Smart Village Program South Asia is particularly vulnerable to climate change and climate variability because of its dense population. The Indian states of Haryana, Punjab and Bihar are located in Northern India on the Indo-Gangetic Plains where 15% of the world’s wheat is produced. Agriculture predominates in these three states with about a 71% net sown area, compared to 43% throughout India. However, water scarcity and soil fertility loss have led to declining yields. Increasingly volatile and unpredictable weather also pose problems for agriculture. As temperatures rise, these problems are expected to worsen (CCAFS 2014). The Consultative Group on International Agricultural Research (CGIAR) Research Program on Climate Change, Agriculture and Food Security and the International Maize and Wheat Improvement Center (CIMMYT) has tried to address these challenges through the ClimateSmart Village Program. This program customizes a package of technologies to the needs of a village to achieve water conserving, soil conserving, energy conserving, weather risk management and greenhouse gas reduction objectives. These packages are designed in conjunction with researchers, farmer cooperatives, the private sector and local government leaders. Technologies and practices promoted include zero tillage, direct seeding for rice, laser land levelling, alternate wetting and drying for rice, residue management, index-based crop 20 insurance, ICT-based weather information and agronomic advice, and precision nutrient management. For this research, we investigated the adoption of three of these technologies: zero till, laser land levelling and direct seeded rice practices (CCAFS 2014). The goal was to estimate the adoption of these three technologies for the three districts (Karnal, Ludhiana and Vaishali) where they have been promoted by national and international research organizations. Groundnut-based farming systems Andhra Pradesh is a semi-arid region in southern India with a population that is heavily dependent on agriculture. The state was previously predominantly rice producing but the last several decades have seen a transition to cash crops including groundnuts. Water scarcity presents a significant constraint. The International Crops Research Center for Semi-Arid Tropics (ICRISAT) has worked in Andhra Pradesh since 1972. ICRISAT’s work in groundnuts in Andhra Pradesh includes introducing a new drought tolerant variety of groundnut, integrated pest management, and sustainable natural resource management, especially water conservation. The groundnut technology adoption survey in the two districts in Andhra Pradesh (Anantapur and Kurnool) measures the adoption of 12 groundnut related technologies and natural resource management practices promoted in the past by ICRISAT and other national partners. These are listed and described briefly below: - Soil Bunds –Within the farm field to collect water runoff and prevent soil erosion. - Field/Boundary Bunds - Border of the farm land to collect water runoff and prevent soil erosion. - Broad Bed and Furrow – System to provide both drainage and standing water 21 - Land Leveling – Reduces slope and conserves water - Contour Bunds - Soil bunds used on sloped lands - Polythene Mulching - Used polythene covers for water conservation - Nala Plugs/RFDs - Dam constructed for save the soil during the rainy season - Sunken Pits – Pits dug to retain water - Farm Ponds – For harvesting rainwater - Masonry Check Dams - Constructed control the water flow and erosion - Well Recharge Pits - For rain water harvesting and saving - Penning Sheep/Goat/Cattle – For compost 3.2 Survey implementers The local enumerator approach was tested through local survey implementers in India. Implementers were identified through a request for proposal issued by MSU that invited responders to propose their own version of the local enumerator approach that was described in the RFP. The proposals received were evaluated on the criteria of innovativeness, rigor, cost effectiveness, and potential for continuation beyond this pilot. Three survey firms were selected based on these criteria, each with a slightly different version of the local enumerator approach. As per the RFP, each implementer proposed to conduct the surveys in two districts based on their experience in carrying out surveys in the past (Table 1). In addition to these three local enumerators, a fourth survey firm was selected to implement the conventional survey in all the five districts. Data from this survey is used for comparison. This fourth survey firm was selected based on a separate expression of interest solicited by MSU. 22 Table 2 summarizes the main differences between the approaches used by the three LEA implementers as well as the one CEA implementer. Table 2: Summary of differences between survey implementation models Firms selected for the pilot study Questionnaire Enumeration software Enumeration hardware Sampling method Firm selected for comparison survey LEA 1 LEA 2 LEA 3 CEA Designed by LEA Designed by LEA Designed by LEA Designed by MSU Implementer 1 with Implementer 2 with Implementer 3 with with input from input from input from input from ICRISAT and ICRISAT and MSU CIMMYT and CIMMYT and CIMMYT MSU MSU Custom androidCustom androidCustom androidSurveyBe based app based app based app Tablet Tablet Tablet and Laptop smartphone Three-stage cluster Three-stage cluster Two-stage cluster Two-stage cluster random sampling random sampling random sampling random sampling 800 600 800 800 Sample size per district (number of households) Sample selection: Stage 1 (Blocks) and 2 (villages based on PPS): By LEA Implementer 1 Stage 3 (10 households/village): By enumerators Enumerator recruitment Local CSO employees Stages 1 (all Blocks) and 2 (villages based on PPS): By LEA Implementer 2 Stage 3 (10 households/village): By enumerators District level agricultural colleges Enumerator training One per district; conducted by LEA Implementer 1; MSU student present as observer One per district; conducted by LEA Implementer 2; MSU student present as observer One per district; conducted by LEA Implementer 3; MSU student present as observer Survey management/ oversight By CSO Survey management at district level; data checked by supervisor By phone 23 Stage 1 (100 villages): By LEA Implementer 3 Stage 2 (8 households/village): By enumerators Stage 1 (80 villages based on PPS): By MSU Stage 2 (10 households/village): By enumerators Local market researchers Organization’s contacts (previous survey participants) One per district; conducted by the CEA Implementer; MSU student present for technical support CEA Implementer staff, mostly by phone; 1-2 initial field visits with MSU student (only in northern districts) LEA Implementer 1 LEA Implementer 1 is a subsidiary of an international economic consulting firm based in Washington DC. LEA Implementer 1 has offices in Delhi and Chennai and provides a range of services including survey design and implementation, data analysis and economic modelling. LEA Implementer 1 conducted the survey of adoption of groundnut-related technologies in Anantapur and Kurnool districts of Andhra Pradesh. Their local enumerator model proposed was to use two local civil society organizations (CSOs) – one in each district – as the source of enumerators. These CSOs do not work directly in agriculture and enumerators mostly have backgrounds in health and the environment.1 Unlike other implementers, LEA Implementer 1 enumerators were compensated through their regular salary received through their CSO employer. In the early stages of the survey, CSO management accompanied enumerators in conducting surveys. They also provided data quality control oversight. All the surveys were conducted in local languages and recorded for verification as part of the CAPI software developed by LEA Implementer 1. LEA Implementer 2 LEA Implementer 2 is a consulting firm based in Delhi that specializes in agricultural development. They carryout qualitative and quantitative research, impact evaluation, project implementation, and extension education and training. Like LEA Implementer 3, their survey looked at farmer adoption of laser land levelling, zero-tillage practices and direct seeded rice 1 Note that because of their regular work in the communities in which they surveyed, Implementer 1 enumerators reported knowing nearly 27% percent of respondents. 24 technologies introduced through the CSVP program in two northern districts of Karnal and Ludhiana. LEA Implementer 2’s approach was to use students from the local agricultural universities, including those who they had worked with before. These enumerators were not from the communities they surveyed and only 0.5% knew the respondents they interviewed. But rather than send a team of enumerators to one village after another, the enumerators were based locally and responsible for interviewing respondents in their nearby area. Data collected by LEA Implementer 2’s enumerators were submitted to a supervisor for approval before final submission to LEA Implementer 2 home office. LEA Implementer 3 LEA Implementer 3 is a consulting firm based in Delhi that provides market research and business advisory services to a range of sectors including agri-business, food processing, infrastructure, logistics, green energy and rural management. LEA Implementer 3 conducted the adoption surveys for three technologies related to the CSVP program: laser land levelling, zerotillage practices and direct seeded rice. Their survey took place in Karnal and Vaishali districts. LEA Implementer 3’s proposed initial approach was to recruit enumerators who were hired from a pool of Community Service Center (CSC) providers. The CSC program is a new initiative of the Government of India (GoI) to make local government provision of services more efficient by outsourcing them to community based private service providers. However, most of the CSC enumerators recruited for this study dropped out because they did not view remuneration as matching the required level of effort by them. LEA Implementer 3 replaced these enumerators with local youth who had experience or education in agriculture. However, data quality was low 25 and the pace of work was slow on account of rice sowing season. Finally, these were replaced by market research experts from the area. LEA Implementer 3 communicated directly with enumerators during the survey to address implementation issues. The two failed attempts at recruiting appropriate local enumerators point to one of the challenges of the approach. Comparison The CEA Implementer is a not-for-profit organization which undertakes surveys, analytical and socio-economic impact evaluations. The CEA Implementer carried out surveys in each of the five districts in which the three local enumerator approach implementers conducted their surveys. The CEA Implementer responded to a solicitation for a conventional survey and conducted the survey mostly following the same survey implementation model they usually follow. Enumerators included some who were hired from the district in which the survey was being carried out as well as from outside the district. Only in Bihar did the CEA Implementer use some enumerators it had worked with previously. One person from the CEA Implementer was assigned to provide management support to all the surveys. CAPI tools All three local enumerator survey implementers developed their own custom android survey applications through different third party software developers. Implementers were aware of off the shelf data collection applications but preferred to develop their own applications. Application development costs in India are extremely low and implementers may have viewed having their own survey application as market advantage. Additionally, one implementer reported having had a bad experience with plot level data collection with an off the shelf data collection application. These applications were developed for tablets which were purchased by each of the three 26 implementers through the project award funding. The surveys they conducted for this research were their first use of these applications and therefore implementers reported some issues during the implementation of the surveys (for e.g. skip logic and GPS location ID), which will likely be resolved in future surveys. The CEA Implementer used laptops for data collection, and the off-the-shelf survey program Surveybe. The laptops were rented and only the rental cost were included in their budget. Unlike the three pilot study firms, they had prior experience in using the CAPI approach and were familiar with the SurveyBe program. 27 4. FRAMEWORK, METHODOLOGY AND DATA 4.1 Framework and Methodology This research seeks to test the local enumerator (LE) approach by testing the three different versions of the approach described earlier. The hypothesis is that the LE approach will produce adoption estimates within the margin of error of a conventional survey, but at lower costs, and thus will be more cost effective. This research empirically tests this hypothesis. The empirical tests involve means comparison (balancing tests) of two main adoption outcome variables controlling for the village and enumerator fixed effects, and some covariates. These tests are based on three pairs of data points, each comparing the local enumerator approach with the conventional survey. Note that while the research tests for statistically significant differences between adoption measurements, it also distinguishes between statistically and practically significant measurement differences. Therefore while using a p-value significance level of 0.05 allows for five percent of comparisons to be different by chance, a larger percentage are expected to be statistically different but practically comparable. That is, while the two approaches may yield statistically significant differences, some of these differences are likely to be close enough so as not to warrant different policy responses. Because of this limited sample size in evaluating the LEA (n = 3), claims about the approach itself are made with caution. Note that this research is only able to make tentative claims about which approach generated better quality data. In this study, the CEA data is not viewed as the “gold standard”; rather it is the comparison group. In other words, the data collected through the 28 CEA is the counterfactual - the default approach conventionally used to conduct representative surveys for tracking technology adoption. In a given district, the sample of farmers selected for data collection by the LEA implementers and the CEA implementer is randomly drawn from the same population. Thus, this research posits that any differences between the LEA and CEA implementer are likely to be due to four factors: sampling differences, enumerator quality differences, questionnaire and data collection application differences and survey management differences. Sampling differences between the two survey approaches may result from the difference in cost structures and lead to differences in the precision (variance) of estimates. The LEA is based on one enumerator per cluster and therefore eliminates inter cluster costs. For the CEA, there are inter cluster costs if the number of clusters exceeds the number of enumerators. Inter-cluster costs comprise of labor for an enumerator and perhaps a driver, per diem and lodging for each, as well as vehicle rental and fuel. Inter-cluster costs imply increased interview costs as the ratio of respondents to clusters decreases. On the other hand, because enumerators are not as mobile in the LEA approach there may be more enumerators and therefore more enumerator fixed costs such as training or data collection technology costs. These differences in cost structures could create incentives for differences in sample designs between the two approaches. In our research, to some extent, sampling differences are controlled through sample weights, village fixed effects and farmer covariates.2 2 In the absence of household lists there also might be selection bias differences if local and conventional differ systematically in how rigorously they follow within village random walk sampling techniques. This cannot be controlled for during analysis. 29 In addition to sampling, data quality is likely to be affected by enumerator’s ability, experience and skills set, which may differ between the LEA and CEA. It is likely that enumerators recruited from the urban areas (as would be the case for CEA) will be characteristically different from enumerators based in rural areas (as would be the case for LEA). This research tests the effects of enumerators on differences in adoption measurement through the inclusion of enumerator fixed effects. It also examines differences in the background between LEA and CEA enumerators, and whether or not these differences account for any differences in the observed adoption outcomes between the LEA and CEA. This allows the research to overcome the limitation of a small LEA and CEA implementer sample size and to determine theoretically whether or not the local enumerator approach should result in lower data quality if on average local enumerators are of lower caliber. Local enumerator implementers had discretion in how they designed questionnaires so long as they collected the minimum required data. This means that the questionnaires used to collect data were not the same for local enumerator implementers and the comparison group. Differences include question wording, skip logic and question location within questionnaire. Because these differences are perfectly correlated with the implementer, we are not able to control for them in this research. Additionally, data collection applications and technologies differed between local enumerator and comparison groups which again could not be controlled for in this research. Questionnaires from each implementer can be found in the Annex. 30 This research also does not explicitly the effects of different survey management practices. These may have an effect on data quality control by incentivizing or enforcing better enumerator performance, or by detecting and effectively addressing potential problems in data collection. Likewise, it does not explicitly test differences in the data collection applications (i.e., the CAPI program and tool) though each enumerator used a different one. Because the sample size of implementers is small (n=3), attribution is difficult, especially given the multi-dimensionality of management practices. However, if, after accounting for sample and enumerator differences there remain statistically significant differences in adoption outcomes, they can implicitly be attributed to differences in implementation practices, including differences in the applications. See Table 2 for a summary of implementation differences across the four implementers. 4.2 Data The surveys were conducted by the four implementers at the end of the rainy season from September to December, 2015. All the data, including technology adoption correspond to the same season and timeframe across the pairs of implementers that implemented the survey within a given district. Quantitative data were collected at both the household and plot level. 31 Table 3: Technologies and adoption outcome variables included in the comparative analysis by implementers Adoption variables Technology Percentage of adopter households Acres of land under adoption per household Wheat based farming system technologies Laser land leveler X X Zero tillage X X Direct seeded rice X X Groundnut based farming system technologies \a Soil bunds X X Field/boundary bunds X X Broad bed and furrow X X Land leveling X X Contour bunds X X Polythene mulching X X Farm ponds X X Masonry check dams X X Well recharge pits X X Penning livestock X X Implementers LEA 1 LEA 2 LEA 3 CEA X X X X X X X X X X X X X X X X X X X X X X X X X X X X X \a Two technologies -- Nala plugs/RFDs and sunken pits are excluded because adoption rates were less than 1%. Household level data are used to estimate two measures of adoption for the comparative analysis: 1) indicator of whether a household is an adopter or non-adopter of a given technology; and 2) total area per household under a given technology or a practice. For both these indicators, adoption refers to the previous or current season (depending on the type of technology and the season in which the focused crop was grown). The list of technologies for which these two adoption variables are calculated and used in the comparative analysis as the key outcome variables are listed in Table 3 by implementers. 32 Given the fact that each implementer designed their own questionnaire (with some feedback from MSU and the CGIAR centers), there is no one-to-one correspondence between the datasets across the four implementers. However, Table 4 lists the variables for which the data collected across implementers are comparable. These variables are used as covariates in the treatment effects estimation models to control for differences in the characteristics of sample across implementers included. Table 4: List of variables common across the four implementer datasets Respondent is head of household Age of head of household Gender of head of household (1= male) Education of head of household LEA Implementer 2/CEA Implementer – Years LEA Implementer 3 – Categorical LEA Implementer 1 – Literacy (y/n) Annual income (1 = <200,000 Rs per hh) Number of people in household Distance to input dealer (km) Household level land owned (acres) Used credit for agriculture (past year) Used crop insurance (past year) 4.3 Sample size For the CEA Implementer surveys, MSU was responsible for determining the sample size and the sampling method. We used a confidence level of 95% and a precision of 5%, to determine the minimum sample size using a simple random sampling (SRS) method. The estimated sample size using SRS was 384 households.3 However, SRS was not feasible logistically and so a cluster 3 This is based on the following standard formula for calculating sample size N = t² * p(1-p) / m², where t is confidence level at 95% (standard value of 1.96), p is estimated parameter value in the project area (50% assumed in this case, which is the most conservative value), and m is the level of precision (assumed 5%). 33 random sampling method was used. This required adjusting SRS for the cluster design effect.4 Using an estimated .12 intra cluster correlation coefficient for adoption and a cluster size of 10 households, the design effect was estimated to be 2.08 which when multiplied by 384 yields the sample size of 800 per district. This information was shared with the LEA implementers, which nudged them to increase their initially proposed sample size which was 400 households per district. Table 5 shows for each district the sample size by implementers, which ranges from 600 households for LEA Implementer 2 conducted surveys to 800 households for all other surveys. The lower sample size for LEA Implementer 2 implies a higher confidence interval of the estimates derived from their sample survey relative to other sample surveys. The sampling method used by each survey firm is described below. Table 5: Sample size by implementer and district District Anantapur Kurnool Karnal Ludhiana Vaishali LEA Implementer 1 Groundnuts (800) Groundnuts (800) LEA Implementer 3 LEA Implementer 2 Wheat based (800) Wheat based (600) Wheat based (600) Wheat based (800) CEA (comparison) Groundnuts (800) Groundnuts (800) Wheat based (800) Wheat based (800) Wheat based (800) 4.4 Sampling method The sample selection for the three LEA implementers was done independently by each implementer. LEA Implementer 1 sampled eight hundred households from within Anantapur and 4 The main components of the design effect are the intra-cluster correlation, and the cluster sample sizes. The design effect (DEFF) is calculated as: DEFF = 1 + d (n – 1), where, d is the intra-cluster correlation for the statistic in question, and n is the average size of the cluster. 34 Kurnool districts. The plan was to select 80 villages from each district and sample 10 households per village. In Anantapur, 41 out of 63 mandals were purposively selected that accounted for 90% of groundnut area in the district. In Kurnool, 22 out of 53 mandals were selected which accounted for 95% of groundnut area sown. This implies that the adoption data are representative of only 90% and 95% of groundnut growing area in these two districts, respectively. For these purposively selected mandals, a two stage cluster sampling method was used. In stage one, the number of villages to be selected from each Mandal was determined based on the probability proportional to size (PPS) method, where size was defined as the area planted to groundnut in each village. To select the determined number of villages, Mandals were divided into quarters geographically and an equal number of villages were selected from each quadrant with villages spaced at least 5 kilometers apart. Within each village, enumerators were instructed to use a random walk technique to select farmer households. This involved identifying the village center and then proceeding to interview households at fixed intervals from the center in different directions. LEA Implementer 2 followed a two-stage cluster random sampling method to select 60 villages across all the blocks in each district. In the first stage, the selection of villages was based on the probability proportionate to size (PPS) method, where the number of villages in each block were determined based on the share of wheat area in the total wheat area planted in the district. Urban areas and villages less than 50 households were excluded from the selection list. In the second stage, 10 households per selected village were identified randomly by the enumerators based on a list of households obtained from the village chief. 35 LEA Implementer 3 sampled 800 households within Karnal and Vaishali districts using a twostage cluster random sampling method. In the first stage, one hundred villages were selected randomly from a list of all the wheat growing villages in the district. In the second stage, within each village, eight farming households were randomly selected from the village’s electoral rolls. For the comparison surveys, a two-stage cluster random sampling method was used in all 5 districts. In each district a list of all the villages was compiled by MSU based on the last census (2010). For sample selection purpose, the few villages where wheat (in Karnal, Ludhiana and Vaishali) or groundnut (in Anatapur and Kurnool) were not identified as one of the top three most important crops were excluded. In the case of Anantapur and Kurnool, only the villages from the mandals selected for the LEA Implementer 1 surveys were included to make the sampling frame comparable to LEA Implementer 1’s. In stage 1, 80 villages were randomly selected by MSU from the list of villages using the PPS method where size was defined as the net sown area (to all the crops) in the village. Within each village, enumerators were responsible for sampling the households as randomly as possible. Enumerators were instructed to purposively visit households in different sections of the village and randomly select 10 households to represent a cross section of village demographics. In order to make the sample estimates from these surveys representative of the population, it is necessary to multiply the data by a sampling weight, or an expansion factor. The sample weights for each household were calculated as the inverse of its probability of selection (calculated by multiplying the probability at each sampling stage). Note that due to problems of village names in the Vaishali data file not matching the list of village level information provided by LEA 36 Implementer 3 to calculate the sample weights, we were not able to calculate the sample weights for the LEA Implementer 3 data from Vaishali. This district is thus excluded from the analysis presented in this paper. 4.5 Costs Total costs are the total contract amount awarded to each implementer. For the CEA Implementer, this includes the costs of MSU researchers who were involved in designing and managing the survey. On the other hand, the three local enumerator implementers received minimal assistance in survey design and management. Costs are broken down into five categories based on actual and not budgeted costs: - Researchers cost – The costs of MSU researchers’ time. - Professional fees - Includes costs of management and supervision as well as overheads charged by each implementer. - Training and survey implementation costs – The costs of training enumerators, payments made to enumerators as well as their travel costs. - Technology – This includes the cost of application development and tablet purchase. For the CEA Implementer, this is rental of enumeration software and laptops. - Other costs – Indirect costs and taxes 37 5. RESULTS 5.1 Comparison of descriptive statistics This section begins with a comparison of descriptive statistics across the three pairs of LEA and CEA implementers across the districts. For LEA Implementer 1, the comparison is for the pooled data across two districts – Anantapur and Kurnool; for LEA Implementer 2, the comparison is for the pooled data across two districts – Karnal and Ludhiana; and for LEA Implementer 3, the comparison is for only one district – Karnal. Household characteristics and adoption estimates Table 6 below shows the weighted mean values for some of the household characteristics for which comparable data were collected across the four implementer surveys. Similarly, Table 7 presents the comparison between the LEA and CEA survey implementers of weighted mean values of the two adoption variables for the technologies focused in this study. Note that in all tables, standard errors are in parenthesis. The main purpose of this mean comparison is to see if the two samples - drawn from the same population but using different approaches - are similar or different based on key adoption variables and covariates. As these results indicate, the two approaches have yielded samples of household that are significantly different across all three implementer approaches. The mean estimates differ significantly at p <0.05 for over half the covariates used for this stud. For LEA Implementer 2 the number of unbalanced covariates is the highest with seven out of ten not matching. Notably large differences include distance to inputs for LEA Implementer 1 and credit use for LEA Implementer 2. 38 Table 6: Comparison of household characteristics between the three pairs of LEA and CEA approaches LEA 1 CEA Difference LEA 2 CEA Difference LEA 3 CEA Difference Demographics N 1,631 47.38 (.68) .59 (.06) .86 (.05) 5.60 (.15) .86 (.03) 1,518 42.79 (.45) .65 (.02) .93 (.01) 4.65 (.07) .90 (.01) 3,149 4.59* (1.14) -.06 (.09) -.06 (0.06) .95* (.21) .04 (.04) 1,188 47.84 (.49) 7.93 (.17) .80 (.02) 6.07 (.10) .60 (.02) 1,540 49.37 (.33) 8.49 (.09) .73 (.14) 5.78 (.07) .23 (.13) 2,728 -1.52* (.72) -.56* (.27) -0.07 (0.04) 0.29 (0.16) .36* (.04) 785 48.98 (.57) 2.36 (.05) .81 (.02) 6.77 (.15) .42 (.02) 729 45.65 (.46) 2.43 (.03) .90 (.01) 6.23 (.13) .27 (.02) 1,514 3.31* (1.11) -.08 (.07) -.10* (.02) 0.53* (.22) .15* (.05) 20.48 (1.04) 16.60 (.32) 4.89* (1.27) 23.10 (.53) 20.99 (.36) 2.10* (.90) 20.72 (.55) 20.45 (.45) .15 (.97) 1626 1513 Land owned 6.29 5.85 (acres) (1.07) (.14) Access to services and infrastructure N 1,591 1,538 Used crop .77 .51 Insurance (.04) (.03) Used credit .98 .67 (.01) (.02) Distance to 10.16 26.55 Inputs (km) (1.24) (1.53) * denotes p <.05 3,139 .44 (.84) 1167 6.96 (.20) 1515 7.21 (.16) 2,682 -.25 (.35) 756 6.54 (.23) 764 7.13 (.22) 1,520 -1.09 (.76) 3,129 .26* (.08) .30* (.06) -16.39* (4.57) 1,148 .07 (.01) .76 (.02) 8.80 (0.21) 1,542 .01 (0) .39 (.01) 7.49 (0.15) 2,690 .06* (.02) .38* (.06) 1.32* (.72) 721 0.15 (0.02) 0.13 (0.01) 9.46 (1.82) 777 0.03 (0.03) 0.18 (0.02) 8.39 (0.18) 1,498 0.13* (0.02) -0.04 (0.03) 1.12 (1.87) Age of HH Head HHH Education Respondent is HHH HH size Low income (1=<200,000 R/year) Years of Farming experience Land holding N The means comparisons in Table 7 also show significant differences in adoption rates and adoption area across all three LEA implementers and their CEA comparison groups. For LEA Implementer 1 the adoption rate results are significantly different for the five groundnut technologies, and for the adoption area variable, results exhibit statistically significant differences between groups for three technologies. Note that nala plugs/RFD and sunken pits both have adoption rates and areas less than .01 and are therefore excluded in all the subsequent analysis. 39 For LEA Implementer 2, the means comparison of the adoption outcome shows two out of three wheat system technologies with statistically significant differences between groups for adoption rate estimates, and one for adoption area estimates. Differences are most pronounced for zero till technology, with area largely a function of the difference in rate. For LEA Implementer 3, five of six adoption indicators are statistically significantly different with the only exception being direct seeded rice area. Similar to LEA Implementer 2 estimates, differences between groups are most pronounced for zero till technology (Table 7). However for several of the statistically significant difference, adoption rates and/or differences are so low that it is questionable whether or not these differences are in fact practically significant. This includes polyurethane mulching adoption rate and area for LEA Implementer 1 and direct seeded rice adoption rates for LEA Implementers 2 & 3. 40 Table 7: Adoption rate and adoption area for different types of technologies: Comparison between the two pairs of survey approaches by implementers Implementer/Technology Adoption Rate LEA LEA Implementer 1 Soil bunds Field bunds Broad bed and furrow Land leveling Contour bunds Polythene mulching Nala plugs/RFDs Sunken pits Farm ponds Masonry check dams Well recharge pits Penning LEA Implementer 2 (N) Laser land leveling Zero Till Direct seeded rice LEA Implementer 3 (N) Laser land leveling Zero Till Direct seeded rice CEA Adoption Area (acres/HH) Differences LEA CEA Differences 1,643 .24 (.04) .18 (.03) .03 (.01) .26 (.05) .13 (.05) 0 (0) 0 (0) 0 (0) .07 (.01) .01 (.01) 0 (0) 0.32 (0.05) 1,536 .28 (.04) .02 (0) .33 (.03) .05 (.01) 0 0 .03 (.02) 0 (0) 0 (0) .03 (.02) .01 (0) .02 (.02) 0.30 (0.03) 3,179 -.04 (.15) .16* (.04) -.30* .04 .21* .05 .12* (.05) .03* (.01) 0 (0) 0 (0) .04 (.03) 0 (0) -.02 (.01) .02 (.06) 1,634 .26 (.04) .18 (.03) .04 (.01) .33 (.12) .18 (.12) 0 (0) 0 (0) 0 (0) .07 (.02) .01 (.01) .01 (0) .59 (.14) 1,526 .38 (.07) .04 (.01) 1.08 (.08) .24 (.05) 0 (0) .09 (.03) (.01) (.01) 0 . .07 (.05) .03 (.01) .02 (.02) .89 (.06) 3,160 -.11 (.19) .14* (.04) -1.05* (.13) .09 (.16) .18 (.11) -.08* (.02) .01 (0) 0 (0) 0 (.05) -.02 (.02) .02 (.01) .30 (.23) 1,541 .55 (.02) .13 (.02) 0 (0) 1,184 .54 (.01) .01 (0) .02 (0) 2,725 .01 (.04) .12* (.03) .02* (.01) 1,176 3.82 (.18) .75 (.11) 0 (0) 1,523 3.11 (.13) .08 (.02) .08 (.04) 2,699 .72 (.38) .68* (.13) .09 (.06) 797 .73 (.02) .22 (.02) 0 (0) 776 .61 (.02) .03 (.01) .04 (.01) 1,573 .12* (.04) .18* (.03) -.04* (.01) 777 4.59 (.22) 1.41 (.16) .02 (.02) 768 2.74 (.12) .18 (.05) .07 (.02) 1,545 1.88* (.34) 1.22* (.22) -.05 (.03) * denotes p <.05 41 Comparison of enumerators’ characteristics Table 8 below shows the comparison of enumerator characteristics for each implementer. This data were collected during enumerator training. The results show that there are significant differences between enumerators recruited to conduct surveys by the LEA implementers and enumerators recruited by the CEA Implementer. However, these differences are not consistent with the expectation that CEA enumerators might be more qualified than local enumerators. In fact, on balance, it is not obvious which group of enumerators are more or less competent for the type of survey they conducted. For example, LEA implementers used smartphones and tablets for interviews and the enumerators recruited by them are slightly more likely to own a smartphone and thus more experienced in using that tool. On the other hand, CEA implementer used laptops and the enumerators reported spending significantly more time using a computer than enumerators from the LEA approach (Table 10). Conventional enumerators have more formal education. However local enumerators have more experience working as enumerators including carrying out agricultural and CAPI surveys. Hours spent per week in a job or other occupation is similar. This pattern holds at the individual implementer level with several exceptions. One is that LEA Implementer 3 enumerators are the only ones with significant CAPI experience. They also all own smartphones. LEA Implementer 2 enumerators have less agricultural experience than their counterparts, a fact that may be accounted for by their younger age. However, they do have more experience conducting agricultural surveys. Only one enumerator was female. 42 Table 8: Comparison of enumerator characteristics Age Education5 Hours/ Ag. Survey Ag. CAPI Owns PC use week in exper. exper. survey (times) smart(hours/ job (years) (times) (times) phone week) 32.4 9.6. 3.4 2.9 .3 .4 6.8 25 .3 .7 .2 .3 .4 32.1 7.4 9.3* 2.7* 2.7* 0 .1 25.4* 10.7 2.8 s.9 .9 .3 .2 5.6 15 (2) 16 (10) 16 (9) 13 (9) 11 (8) 16 (9) 16 (10) LEA 1 CEA Diff. SE n (control) 34.4 24.2 10.3* 1.7 16 (11) 7.9 8.4 .5 .3 16 (11) LEA 2 CEA Diff. SE n (control) 25.4 28.1 2.7 3.2 12 (8) 6.8 8.8 1.9* .4 12 (9) 20.2 26.1 5.9 5.2 11 (9) 3.2 6.2 3.1 2.3 12 (9) 2.1 .4 1.6 .8 12 (9) .9 .4 .5 .5 12 (9) .1 .1 0 .1 12 (9) .6 .4 .1 .2 12 (9) 20 35.2 15.2* 6.8 11 (9) LEA 3 CEA Diff. SE n (control) 25.3 26 .8 1.1 8 (3) 7.6 9 1.4* .3 8 (4) 30.3 24.8 5.5 3.2 8 (4) 8.9 4.5 4.4* 1.7 8 (4) 8.4 0 8.4* 1.5 8 (4) 5.4 0 5.4* 1.5 8 (4) 4.5 0 4.5 2 8 (4) 1 0 1* 0 8 (4) 20.4 18 2.4 8 (1) All LE 29.4 CEA 25.8 Diff. 3.5 SE 2.2 N (control) 36 (19) 7.5 8.6 1.1* .3 36 (20) 27.9 25.9 2 4.3 34 (11) 7.3 3.1 4.2* 1.8 36 (19) 4.1 .6 3.5* .8 36 (18) 2.8 .3 2.5* .7 33 (18) 1.3 .2 1.1 .7 31 (17) .6 .4 .1 .1 36 (18) 14 33.3 19.2* 4.1 35 (16) * denotes p <.05 5.2 Balancing tests for covariates Because the means comparisons noted above revealed significant differences between group characteristics, balancing tests are conducted on the household and farm characteristic variables (i.e., covariates) by including village fixed effects in the regression model below. Village fixed effects account for village level differences in the two samples that may affect differences in household characteristics. 5 1=Illiterate, 2=Read & Write (Non formal Education), 3=Primary (1st- 5th), 4=Upper Primary (6th-7th), 5=High School (8th 10th), 6=Higher Secondary (11th-12th), 7=Diploma/ ITI, 8=UG, 9=PG & Above 43 𝑐 = 𝛽𝑥𝑖 + 𝛾𝐯 + 𝜇 for i=1, 2, 3 (3) Where c is the covariate variable, xi is an indicator for implementer i, v is a vector of villagelevel indicator variables, and u is the error term. The balancing test in model 3 is conducted separately for each of the three LEA implementers. In each case, variable x=0 represents data collected by the CEA Implementer (i.e., the comparison data) in the same districts. The estimate of interest is 𝛽, which measures the difference in the estimated value of variable c in the sample data collected using the LEA approach compared to the estimated value of the same variable derived from the sample data collected using the CEA approach (the control group). Model 3 is repeated for all ten covariates with all three pairs of LEA-CEA implementers and the results (i.e., 𝛽 coefficients) are shown in Table 9. Note that in all the balancing tests conducted using the regression based approach, the model includes sample weights, and robust standard errors which are clustered at the village level. Village level clustering of standard errors accounts for hypothesized greater correlation within villages than between them. For LEA Implementer 1, after accounting for the village fixed effects, four out of ten covariates are still unbalanced (Table 9). For LEA Implementer 2, three covariates remain unbalanced compared to seven without fixed effects. Finally, three out of 10 LEA Implementer 3 covariates are unbalanced compared with five using simple weighted means comparisons. While the sample is now more balanced than noted in Table 6, there are still significant number of covariates that are unbalanced between the two samples, indicating that controlling only for the differences in the villages included in the two samples is not adequate to account for the differences in the 44 observed estimates of the adoption variables. This implies that any analysis of the treatment effect to evaluate the effectiveness of the LEA approach in generating adoption estimates comparable to the CEA approach (which is used as a control in this study), must control for differences in the covariates that can influence the adoption outcome variables. Table 9: Balancing tests for covariates with village fixed effects Demographics (n) Age off HHH Literate HHH or HHH Education* Respondent is HHH Number in HH Low income (<200,000 R/year) Years farming Land (n) Land owned (acres) Access (n) Used crop insurance Used credit Distance to input Dealer (km) LEA 1 3,119 6.66* (1.28) -0.26 (0.17) -0.35* (0.11) 0.22 (0.38) 0.02 (0.04) 2.84* (0.88) 3,110 4.60* (1.75) 3,095 0.21 (0.15) 0.16 (0.12) -4.95 (6.78) LEA 2 2,722 -1.40 (2.17) -1.206 (0.69) -0.06 (0.07) -0.59 (0.31) 0.43* (0.05) 0.47 (2.23) 2,670 -1.16 (0.61) 2,677 0.03 (0.02) 0.41* (0.17) 5.45* (2.62) LEA 3 1,530 3.74 (2.50) -0.217 (0.12) -0.11* (0.03) 0.65 (0.48) 0.22* (0.07) 0.19 (2.05) 1,490 -1.09 (0.76) 1,474 0.14* (0.04) -0.07 (0.04) -1.55 (0.92) * denotes p <.05 5.3 Identifying adoption differences After determining that the sample characteristics are unbalanced even with village fixed effects, the next step in the analysis is to identify whether or not there are statistically signficant 6 7 n = 2,636 n = 1,489 45 differences in adoption estimates between LEA and CEA data. For this analysis we use two simple linear regression models. One includes the village fixed effects (v) and the covariates (c) listed in Table 8 (model 4), and the other replaces village fixed effects with enumerator fixed effects (e) (model 5). 𝑦 = 𝛽𝑥 + 𝛾𝒗 + 𝜹𝒄 + 𝜇 (4) 𝑦 = 𝛽𝑥 + 𝛾𝒗 + 𝜽𝒆 + 𝜇 (5) Variable y represents the adoption outcome (i.e., adoption indicator and adoption area per household). Variable x is the indicator of whether the dataset was collected using the LEA approach (x=1) or using the CEA approach (x=0). The coefficient of this treatment variable, 𝛽, measures the difference in the value of the estimated adoption variable from the two approaches, after controlling for the confounding factors included in the models. The hypothesis we want to test is that 𝛽=0. In other words, we hypothesize that the estimates of adoption rate and level of adoption derived from the LEA approach are not statistically significantly different from the estimates derived from the CEA approach. The results from models 4 and 5 are presented in Table 10. The first column simply replicates the weighted means comparisons from the descriptive statistics section (Table 7). The second column shows the results from equation 4 which includes village fixed effects and covariates. The third column 3 replaces village fixed effects with enumerator fixed effects (equation 5). 46 Table 10: Differences in the mean estimates of adoption variables derived from LEA and CEA datasets: Comparison of results across three methods Adoption Rate (A) (B) (C) Means Village FE + Enum. FE + comparison Covariates Covariates LEA 1 (n) Soil bunds Field bunds Broad bed and furrow Land Leveling Contour Bunds Polythene mulching Farm ponds Masonry check dams Well recharge pits Penning LEA 2 (n) Laser land Leveling Zero Till DSR LEA 3 (n) Laser land leveling Zero Till DSR (A) Means comparison Adoption Area (B) (C) Village FE + Enum. FE + Covariates Covariates 3,179 -.04 (.15) .16* (.04) -.30* .04 .21* .05 .12* (.05) .03* (.01) .04 (.03) 0 (0) -.02 (.01) .02 (.06) 3,000 -.58* (.26) .06* (.03) -.43* (.05) .21* (.04) .20* (.04) -.07* (.01) -.07* (.02) -.01 (.01) -.07* (.02) .03 (.12) 2,973 .85* (.08) .60* (.10) -.27* (.07) .23* (.08) .09* (.04) 0 (0) .44* (.12) -.12* (.06) 0 (.01) -.42* (.09) 3,160 -.11 (.19) .14* (.04) -1.05* (.13) .09 (.16) .18 (.11) -.08* (.02) 0 (.05) -.02 (.02) .02 (.01) .30 (.23) 3,000 -.89* (.24) -.02 (.05) -1.20* (.33) .37* (.13) .48* (.05) -.11* (.01) -.26* (.05) -.08 (.09) -.08* (.01) -.05 (.25) 2,971 .43* (.14) -.04 (.20) -1.13* (.32) -.11 (.23) .13* (.06) .01 (.04) -.18 (.22) -.59 (.32) .03 (.02) -2.22* (.51) 2,725 .01 (.04) .12* (.03) .02* (.01) 2,533 .03 (.11) .11* (.04) -.03 (.02) 2,517 .27* (.07) .03 (.03) 0 (0) 2,699 .72 (.38) .68* (.13) .09 (.06) 2,512 1.04 (1.21) .53 (.31) -.08 (.06) 2,507 2.18* (.66) .27 (.27) 0 (.01) 1,573 .12* (.04) .18* (.03) -.04* (.01) 1,352 .12 (.06) .16* (.04) -.06* (.02) 1,352 .07 (.07) .16* (.05) 0 (0) 1,545 1.88* (.34) 1.22* (.22) -.05 (.03) 1,330 2.31* (.76) 1.09* (.30) -.10* (.05) 1,330 .99* (.48) 1.37* (.46) -.01 (.01) * denotes p <.05 47 The comparison of differences across the three methods shows that only for DSR area for LEA Implementer 2 are adoption results statistically indistinguishable across all three models. On the other hand, broad bed and furrow shows statistically significant difference in estimates for both adoption rate and adoption area across all three methods. The number of significant differences varies between these two extremes for the other technologies. In general, the inclusion of village fixed effects and covariates (column B) makes the effect size become statistically significant for more number of technologies compared to simple mean comparison (column A). Replacing village fixed effects with enumerator fixed effects (column C) has only a small effect on reducing the number of statistically significant coefficients compared with the preceding model (column B). Again it is important to keep in mind the distinction between statistical and practical significance. While polythene mulching and DSR both exhibit statistically significant differences across estimation models, these differences are minimal and not likely practically significant in terms of policy decision making. The results indicate that adding village fixed effects and covariates does not generally decrease the number significant differences between adoption estimates. This suggests that while there are differences in sample composition, these differences do not explain the differences in adoption estimates. However, when enumerator fixed effects are included, the number of significant differences in adoption estimates is slightly reduced. The number of adoption rate differences declines from eleven to ten whereas the number of adoption area differences declines from 10 to seven across all three LEA implementers. This indicates that differences between enumerators may indeed explain some of the differences in adoption estimates. The proceeding section explores this relationship. 48 5.4 Comparison of missing values One proxy for data quality are missing values which are themselves bad data but also potentially evidence of more bad data. This research examines the number of missing values for the ten covariates across the four implementers.8 The mean of the total number of missing values per respondent across the ten covariates included in the previous models is compared at the implementer level, as well as at the approach level (Table 11). Households that were interviewed by the local enumerator approach had on average .06 total more missing values out of the ten covariate questions included in this analysis. This difference is highest for LEA Implementer 3 and lowest for LEA Implementer 2. Table 11: Average number of missing values (out of ten covariates) per respondent LEA 1 (1,643) .10 (.04) Control (1,540) .03 (.01) LEA 2 (1,195) .10 (.02) Control (1,542) .08 (0) LEA 3 (801) .27 (.03) Control (777) .11 (.02) All (3,639) .12 (.02) Control (3,082) .05 (.01) Additionally, village fixed effects are included as shown in equation 6 below, where m represents number of missing values. This analysis is also conducted at the implementer level using block level fixed effects.9 𝑚 = 𝛽𝑥 + 𝜌𝑟 + 𝛾𝒗 + 𝜇 (6) Additionally, the model is estimated by including the enumerator FE (equation 7). Both models include an indicator variable for whether or not the respondent is the HH head (r). 8 Due to data issues missing values could not be computed for adoption. Stata variable limits were exceeded with village FE model. Thus, we used the next level of administrative boundaries (blocks, which are equivalent to counties). 9 49 𝑚 = 𝛽𝑥 + 𝜌𝑟 + 𝜽𝒆 + 𝜇 (7) Table 12 shows group comparisons of means as well as regression coefficients estimated using models 6 and 7. The regression results show significant changes in missing values with the inclusion of enumerator fixed effects but not village fixed effects. In fact the model with enumerator fixed effects shows no significant differences in missing values between all LEA and CEA data implying that differences between enumerators account for differences in missing values. Whether or not the respondent is the head of the household in all cases but one has a negative coefficient, indicating the importance of interviewing the head of household, Table 12: Missing value differences and coefficient comparison Mean difference LEA Implementer 1 (n=3,115) Village FE .06 (04) .04 (.04) .02 (.02) .01 (.04) -.07 (.04) .18 (.11) -.08* (.04) .06* (.02) -.08* (.03) HHH is respondent LEA Implementer 2 (n=2,718) .02 (.03) HHH is respondent LEA Implementer 3 (n=1,544) .16* (.07) HHH is respondent All LEA (n = 6,610) .07* (.03) HHH is respondent Enumerator FE -.07* (.02) -.18 (.14) .10 (.06) -.06* (.03) .09 (.09) -.11* (.04) .02 (.02) -.12 (.06) * denotes p <.05 5.5 Exploring the factors contributing to differences in data quality Given the differences in enumerator characteristics noted above, the next step is to see how these characteristics affect the data quality as reflected in the estimated adoption variables and the 50 number of missing values. To do this, enumerator characteristics k are regressed on total adoption area (z) across all 10 technologies for LEA Implementer 1 and all three technologies for LEA Implementer 2 and LEA Implementer 3 (equation 8). Because of missing values, the characteristics included in the model are restricted to enumerator education, experience in agriculture, experience as an enumerator, ownership of a smartphone and ownership of a computer. Additionally, the equation includes an indicator variable for whether or not the respondent is the head of the HH (r). This attempts to control for the respondents’ own knowledge of their farm’s adoption outcomes. As usual, the model include village FE and standard errors are clustered at the village level. 𝑧 = 𝛽𝑥 + 𝜌𝑟 + 𝝉𝑘 + 𝜸𝒗 + 𝜇 (8) The coefficients for vector k and their standard errors are reported in Table 13. The only implementer for which enumerator characteristics are significant in affecting the estimated total adoption area is LEA Implementer 1. The variables that are found to significantly impact the estimated total adoption area are enumerator’s education and whether he/she owns a computer. However given that these are not significant for the other two implementers, it is difficult to draw conclusions on the meaningfulness of enumerator characteristics for adoption area measurement. 51 Table 13: Enumerator characteristics coefficients with total technology adoption area dependent variable Education LEA 1 n =2,356 LEA 2 n = 2,718 LEA 3 n = 1,544 1.27* (.50) .15 (.44) 1.20 (1.52) Enumerator characteristics Agricultural Enumeration Owns Owns PC experience experience smartphone (years) (times) -.15 .11 2.11 4.42* (.11) (.26) (1.11) (.72) -.08 -.43 1.04 -.03 .07 (.25) (.92) (.59) .05 .50 1.10 -1.32 (.36) (.34) (3.97) (1.65) * denotes p <.05 Next the same regression is repeated replacing total adoption area (z) as the dependent variable with m (number of missing values across 10 covariates). In total both enumeration experience and owning a smartphone are significant, with each having a counterintuitively positive affect on the number of missing values. Very speculatively, it could be that these practical skills – enumeration experience and technology familiarity make enumerators more likely to record the true unknown responses of a respondent rather than fabricate a response. Table 14: Enumerator characteristics coefficients with missing values dependent variable Education LEA Implementer 1 n =2,356 LEA Implementer 2 n = 2,718 LEA Implementer 3 n = 1,544 All n = 5,851 .02 (.02) .04 (.03) -.47* (.19) -.02 (-.02) Agricultural Enumeration Owns Owns PC experience experience smartphone (years) (times) -.01* .03* .08 (0) (.01) (.07) 0 .01 .17 (.01) (.02) (.12) -.06* .03 Omitted (.02) (.03) 0 .02* .04* (0) (.01) (.02) * denotes p <.05 52 .01 (.01) .04 (.04) .29 (.13) -.04 (.02) 5.6 Cost The following table shows survey costs and costs per respondent for all four implementers. The average cost per respondent for the three local enumerator survey implementers is $31 dollars, which exceeds the conventional survey (CEA) cost by nine dollars per respondent. However, several factors should be noted that can potentially change this comparison. One is that the cost of developing the Android based data collection application and purchasing tablets together account for about 20 percent of the total cost for the three local enumerator implementers. This is a one-time cost that will decrease as a share of per respondent costs with subsequent surveys. The CEA Implementer on the other hand used an off-the-shelf program, Surveybe, that requires annual subscription and rented laptops, making these costs reoccurring for them. The cost of these two items is substantially lower than the costs incurred by the LEA implementers for the CAPI survey. Taxes also account for some of the difference. Because of its non-profit status, the CEA Implementer did not pay the 14 percent tax paid by LEA Implementers 2 & 3. LEA Implementer 1 also did not pay this taxed based on its international subsidiary status. Table 15: Actual costs for survey implementation (USD) Researchers’ cost (MSU personnel) Professional fees (survey management & oversight) Training and survey cost (enumerator payment, travel) Technology (application and hardware) Overhead and taxes Total cost Total respondents Costs per respondent LEA 1 LEA 2 LEA 3 All LEA CEA 5,000 5,000 5,000 29,500 (9%) (9%) (20%) (11%) (33%) 21,754 22,226 3,790 17,655 (40%) (40%) (15%) (36%) (20%) 20,967 12,129 6,113 31,692 (39%) (22%) (25%) (29%) (36%) 3,871 8,065 7.258 7,950 (7%) (14%) (30%) (14%) (9%) 2,576 8,252 2,403 1,280 (5%) (15%) (10%) (10%) (1%) 48,169 51,681 19,564 88,079 1,600 1,200 1,600 4,000 34 46 15 31 22 53 Accounting for multiple use of technology and controlling for tax differences brings the comparison closer. Specifically projecting over ten surveys (dividing technology costs by ten for LEA implementers) and adding 14 percent tax on the total survey cost less researchers cost makes the LEA average per-respondent cost 28 dollars compared to 24 dollars for the CEA Implementer. Additionally, there are survey implementation economies of scale associated with the larger sample size such as developing the questionnaire and training materials. This gave the CEA Implementer, who implemented their surveys across 2.5 times more households than any local enumerator implementer. Finally there is reason to believe that the CEA may have under budgeted with a price much lower than the competition. This is supported by the fact that all three local enumerator implementers stated that the LEA they used in this study implied close to fifty-percent cost savings for them compared with their conventional, paper-based surveys. 54 6. DISCUSSION AND CONCLUSION 6.1 Summary of findings This research sought to test whether or not the local enumerator approach – using locally-based enumerators and tablets - is a feasible alternative to the conventional enumerator approach. That is whether it would produce statistically indistinguishable adoption estimates at a lower cost. To do this, agricultural technology adoption data were collected by three survey firms in India utilizing different interpretations of the local enumerator approach. This was compared with similar data collected through a conventional enumerator model for the same representative geographies. Through descriptive statistics, the research identified statistically significant differences in the adoption rate and area measurements and also found that covariates hypothesized to influence adoption – such as credit access and land size – were statistically significantly different. The research then attempted to more rigorously identify the difference in adoption measurement between approaches through two strategies. The first strategy was to control for village level differences in samples through village fixed effects and to control for covariates such as land size and access to credit to control for differences between them that might affect adoption estimates. This strategy did not reduce the measurement differences in adoption outcomes. The second empirical strategy was to include enumerator fixed effects in place of village fixed effects to control for differences in enumerators. This did slightly reduce the number of significant differences in adoption outcomes. Since there still remained statistically significant differences even after controlling for the sample composition and enumerator FE, it was not possible to 55 conclude that the local enumerator approach, as implemented through this research, yielded the same adoption measurement results as the local enumerator approach. The fact that there are differences in measurement of adoption does not necessarily imply that the CEA produced the more accurate data. One proxy for data quality is the number of missing values. The CEA did in fact produce .07 fewer missing values on average for the ten responses tested per respondent. However, this was sensitive to model specification and in fact disappeared when controlling for enumerator fixed effects and whether or not the respondent was head of household. Finally, the research found differences in enumerator characteristics between CEA and LEA, though it is not intuitive which group of enumerators would favor the observed differences. Regression analysis was used to assess the effects of enumerator characteristics on adoption measurement and data quality. The results of these analyses were ambiguous and did not help clarify what factors were contributing to differences in adoption estimates and data quality. Before undertaking this research, it was speculated that the local enumerator approach would be lower cost because of reduced transport and per diem costs. The results of the cost comparison are non-conclusive. On average, the LEA came out nine dollars more expensive per HH sample than the CEA. This gap is cut in half when controlling for the fixed cost of technology purchased by LEA Implementers as well as difference in taxes charged. There is also significant variation in costs between the three LEA implementers with the lowest cost 15 dollars per household and the highest 45 dollars per household. 56 A key finding of this research is the significance of the respondent being the head of the household in reducing missing values. This intuitive finding implies that the local enumerator approach has an advantage because of the comparative ease of revisiting households if the head of household is not initially available. 6.2 Sources of differences in adoption estimates To guide future research, it is important to highlight and speculate about the sources of differences in adoption measurement that the analysis was unable to answer. One potential source of difference not studied in this paper is the difference in level of training given to enumerators on each technology focused by the survey, and how to communicate to the farmers what these technologies were. Because many of the technologies are not actual material technologies (i.e., inputs or products one can purchase regularly from the market) but in fact agronomic and resource management practices, proper description of these practices to farmers requires more knowledge and communication skills than for soliciting responses for simpler technologies. Differences in how implementers trained enumerators on these technologies, how well enumerators learned and understood these technologies or already knew about them, and how skilled they were at communicating them to farmers in local languages might have had an effect on adoption measurement. The technologies were not described on the questionnaire so this cannot be evaluated ex-post. Additionally, as is discussed below, the analysis does not control for differences in data collection applications that may be correlated with differences in seeking and recording responses. 57 Differences in questionnaires between implementers may also have affected differences in responses (Kasprzyk 2006). This includes differences in question phrasing, question sequencing, response options and skip logic and the overall length of the instrument. A closely related potential source of difference is differences between the data collection applications. Unlike what Caeyers et al. (2012) found, there were still a significant number of missing values in all of the datasets received, especially for the dataset received from the LEA implementers. This is somewhat to be expected as they each developed their own application, programed them with questionnaires and used it for the first time with these surveys. As the applications are improved and the implementers become more experienced in using them, this source of bad quality data is likely to decrease. On the other hand, the CEA Implementer used an off the shelf application that they had already had experience using and programing. This difference in technology maturity and familiarity with the technologies could possibly account for differences in results seen here. Contrary to expectations, local enumerators were not clearly less qualified than conventional enumerators at least according to the characteristics collected from them. This was initially hypothesized to be the primary potential source in measurement differences for the local enumerator approach. However, caution should be used in extrapolating this finding to other developing country contexts as India’s rural population is likely more educated than other developing countries’. As discussed previously the costs of the CEA Implementer are likely lower per respondent due to the much larger scale of their surveys resulting in design and management efficiencies. Furthermore, there are reasons to expect the local enumerator cost per household to decrease 58 over time. One reason is that the initial cost of software development will decrease as a share of survey cost with each subsequent survey. It is also possible that with regular data collection, the comparative costs of LEA data collection will decrease because of less enumerator attrition. Future research could focus on these sources of differences that this research was unable to address. Most useful would be identifying the characteristics associated with good enumerators through research that unlike this research controlled for approach, implementer, questionnaire, CAPI and other differences. This would have broad implications beyond the local enumerator approach and to date little research has been carried out on the topic. 6.3. Feasibility of the local enumerator approach The local enumerator approach was envisioned to be a commercially sustainable model for collecting relatively simple agricultural data. The idea behind proposing this approach was that an infrastructure of local enumerators could be established and expanded that could be called upon to carryout surveys in their areas on an ongoing basis. Unfortunately, this research was not able test the LEA with longitudinal data collection where it should theoretically hold the most advantage on cost due to reduced travel to and from survey location. Ideally the local enumerator approach would be tested with longitudinal data collection. Deployment for a project’s regular monitoring and evaluation is one example of how the approach could be tested longitudinally. While this research did not directly test this commercial feasibility, the three implementers provided their feedback on the viability of the proposed model. The key lesson learned from this pilot study is that there must be frequent enough surveys with the same implementers for the 59 enumerators to maintain the survey implementer-local enumerator relationship. Speculatively, local enumerators should be engaged in data collection at least twice a year to maintain a relationship with their survey firm. This requires entrepreneurship from survey firms to secure sufficient clientele base in advance to meet this frequency requirement. Furthermore, number of interviews per enumerator is an important parameter for the long-term success of the local enumerator model. On one hand, tablets and training are fixed costs and so data collection costs increase with the number of enumerators per survey. On the other hand, some potential enumerators were discouraged from participating because of the extensive local travel while others complained of travel time after going through with the survey. One possibility for commercialization of the local enumerator approach is in implementing regular project monitoring and evaluation activities that would guarantee bi-annual or quarterly data collection over a multiple year period. Another, which several implementers pointed out, is to sell data to commercial clients on a regular basis. One suggestion was to identify demand for tractor rental services and sell it to mechanized service providers. Another suggested idea for the application of the LEA was to track production of crops and sell that data to traders and processors so they could better plan their purchasing. Similarly, data could be sold to input suppliers. A key challenge here is to identify a sample and local enumerator-centered clusters that meet the sample needs of a broad number of clients. This is likely to be especially challenging during early stages of implementing the model when the enumerator infrastructure is smaller. Like most infrastructure, a certain scale is required to achieve commercial viability. 60 REFERENCES 61 REFERENCES Brogan, D. 2005. Sampling error estimation for survey data. Chapter in Household Sample Surveys in Developing and Transition Countries. United Nations, New York. Caeyers, B., Chalmers, N. & De Weerdt, J. 2012. Improving consumption measurement and other survey data through CAPI: Evidence from a randomized experiment. Journal of Development Economics 98: 19-33. Carletto, C., Gourlay, S, & Winters, P. (2013). From guesstimates to GPStimates: land area measurement and implications for agricultural analysis. Policy Research Working Paper, 6550. World Bank, Washington, DC. Carletto, C., Savastano, S. & Zezza, A. 2011. Fact or Artefact: The Impact of Measurement Errors on the Farm Size - Productivity Relationship. World Bank Policy Research Working Paper No. 5908. Chao, L.W., Szrek, H., Peltzer, K., Ramlagan, S., Fleming, P., Leite, R., Magerman, J., Ngwenya, G.B., Pereira, N.S. & Behrman, J. 2012. A comparison of EPI sampling, probability sampling, and compact segment sampling methods for micro and small enterprises. Journal of Development Economics 98: 94-107. CGIAR Research Program on Climate Change, Agriculture and Food Security (CCAFS). (2014). Climate-smart villages in Haryana, India. http://ccafs.cgiar.org/publications/climate-smartvillages-haryana-india Doss, C. 2006. Analyzing Technology Adoption: Challenges and Limitations of Micro Studies. Agricultural Economics 34(3): 207-219. Dotchin, et al. (2007). The prevalence of Parkinson’s disease in rural Tanzania. Movement Disorders, 23(11), 1567-1572. E-agriculture. (2012). ICT for data collection and monitoring and evaluation. E-sourcebook, Forum 3. http://www.fao.org/docrep/017/aq003e/aq003e.pdf Feder, G. & Umali, D. 1993. The adoption of Agricultural Innovations: A Review. Technological Forecasting and Social Change 43: 215-239. Government of India (GOI) a: Office of the Registrar General & Census Commissioner. (2011). Census Operations, New Delhi. http://censusindia.gov.in/Data_Products/Library/Indian_perceptive_link/Census_Operation_lin k/censusoperation.htm 62 Government of India (GOI) b. 2011. Overview of Census 2011. http://censusindia.gov.in/2011prov-results/data_files/mp/02Introduction.pdf Hausman, J. 2001. Measured Variables in Econometric Analysis: Problems from the Right and Problems from the Left. Journal of Economic Perspectives 15(4): 57-67. Jack, B. Kelsey. 2013. Constraints on the adoption of agricultural technologies in developing countries. Literature review, Agricultural Technology Adoption Initiative, J-PAL (MIT) and CEGA (UC Berkeley). Kasprzyk, D. 2005. Measurement error in household surveys: sources and measurement. Chapter in Household Sample Surveys in Developing and Transition Countries. United Nations, New York. McKenzie, D. & Rosenzweig, M. 2012. Preface for symposium on measurement and survey development. Journal of Development Economics 98: 1-2. Private discussion. 2014. Discussion with key informant. Royce, D. 1986. Address register research for the 1991 census of Canada. Journal of Official Statistics, 2(4): 447-455. Sitati, N.W., Walpole, M.J. & Leader-Williams, N. 2005. Factors affecting susceptibility of farms to crop raiding by African elephants: using a predictive model to mitigate conflict. Journal of Applied Ecology 42(6): 1175-1182. Smith, G., Pellissery, S., Rajan, S. & Dubuc, S. 2007. India microdata scoping study: Final report to the ESRC. Department of Social Policy and Social Work, Oxford. Statistical Policy Office (SPO), US Government. 2001. Measuring and Reporting Sources of Error in Surveys. Statistical Policy Working Paper 31.Washington, DC, USA. United States Agency for International Development (USAID). 2012. Mobile applications for monitoring and evaluation in agriculture. Briefing paper. Washington, DC, USA. Udry, C. 1991. Rural Credit in Northern Nigeria. PhD Dissertation, Yale University. United States Agency for International Development (USAID). 2011. Community knowledge worker. ICT and AG profile. Washington, DC, USA. The World Bank. 2015. Agriculture and Rural Development Dataset. Accessed at http://data.worldbank.org/topic/agriculture-and-rural-development. 63 The World Bank. 2015. World Development Indicators: Rural environment and land use. Accessed online at http://wdi.worldbank.org/table/3.1 Ziliak, S. & McCloskey, D. 2008. The Cult of Statistical Significance: How the Standard Error Costs Us Jobs, Justice and Lives. University of Michigan Press, Michigan. 64 APPENDICES 65 APPENDIX A: LEA Implementer 1 Questionnaire10 SECTION A: GENERAL INFORMATION – ENUMERATOR Date of Survey A1. Name of the Enumerator (DD/MM/YY YY) A3. Place of residence of the Enumerator A2. Name of the NGO 1=MASS, 2=CHRD A4. Mode of transportation used by the enumerator to come to the place of this interview Start time of interview A5. Time required to travel from his/her location to this village/hamlet (HH-MM) 1=Bus, 2=Train, 3=Two Wheeler, 4=Bicycle, 5=Auto, 6=By walk, 99=Others (Specify) GENERAL INFORMATION - LOCATION A6. State A7. District A8. Mandal A10. A11. Enumerator’s acquaintance with farmer Village/Ha GPS Coordinates 1=Yes-in the context of another survey, 2=Yes-in a social context, 3=Yes-we are personal acquaintances, 4=Not met before mlet SECTION B: GENERAL HOUSEHOLD INFORMATION B1. Name of B2. Are you the B3. If the respondent is not the Head of the HH, Ask the main Head of the HH? (If his/her relationship with the Head decision B4. HH Sl. No yes go to B4;If no 1=Spouse, 2=Son, 3=Daughter, 4=Grandson, 5=Granddaughter 6=Brother, 7=Son-in-law, 8=Daughter-in- law, 9=Father, 10=Mother, 99-Others maker in the go next question) (specify) (1=Yes, 2=No) household 10 Introduction and consent removed for brevity 66 A9. Gram Panchayat B5. Decision maker’s age B6. Sex (1=Male, 2Female) B7. Religion of the household 1=Hindu, 2=Muslim, 3=Christian, 4=Sikh, 5=Buddhist, 98=Don’t know, 99=Others (specify) B9. Main Occupation 1=Agriculture, 2=Livestock/poultry keeping (inc. sale), 3=Trading Livestock/poultry & its products, 4=Business, 5=Salaried, 6=Agri/Non-agri. labour, 7=Housewife, 8=Student, 9=Not applicable, 99=Others B8. Marital status of the main decision maker (1=Married, 2=Single, 99=Others) B9a. Main decision maker B9b. Spouse B11. Mobile Number of the Main decision maker B12. Total members in the family B14. Total Annual Income of the family (combined of all working members) (in Rupees) B15. Total Income from groundnut farming? (in Rupees) B10. Literacy Status 1=Illiterate, 2=Read & Write (Non formal Education), 3=Primary (1st- 5th), 4=Upper Primary (6th-7th), 5=High School (8th -10th), 6=Higher Secondary (11th-12th), 7=Diploma/ ITI, 8=UG, 9=PG & Above, 10=Not applicable B10a. Main Decision maker B10b. Spouse B13. Total working members in the family (in Numbers) B13.a. Male B16. What type of Ration card do you have? 1=White Color, 2=Pink Color, 3= Yellow Color, 5=Do not have ration card, 99=Others (Specify) B13.b. Female B17. Do you have a bank account? (1=Yes, 2= No) B18. Does anyone in your household belong to a farmer producer organization or a farm cooperative? (1=Yes, 2= No) SECTION C: LAND HOLDINGS, USES AND GROUNDNUT CULTIVATION DELETED DELETED C1. What is the total C2. In the past two years, how many times have you lost a operational land more than 20% portion of your crop production due to (acres) holding of unexpected weather (e.g., low rainfall, flooding, unexpected your household’s in monsoon time, hail, etc)? the past 12 months? 67 (in Numbers) C3. Number of years you have been producing groundnuts C4. Have you accessed credit for agricultural production in the past 12 months (1=Yes, C5. If YES, from where? C6. If NO, why not? 1=Trader, 2=Local money lender, 3=Bank, 4= Co-operative Bank, 5=Neighbor, 6=Family member, 7=Friends and Relatives, 8=Other financial institutions, 99= Others (Specify) 1= Did not need, 2= Do not have access to credit, 3=Very high interest rate , 99= Others (Specify) 2= No) C7. Do you insure your crops (i.e., do you have crop insurance policy)? C. 7a. If NO, why not? 1=Do not need, 2=Not available in my place, 3= Not available for groundnut, 4=Too expensive, 99= Others (Specify (1=Yes, 2= No) C8. During the past year, where did you receive most of your information and advice about groundnut production and marketing from? (Tick all that apply) 1=Extension agent, 2=NGO staff, 3=Trader/input dealer, 4=Farmer group/leader farmer, 5=Service Provider, 6=I did not receive any information or advice, 99= Others (specify) C8. a. If YES in previous question, In total, approximately how many times did you receive information last year from all these sources? 1=1-2 times, 2=3- C9. Do you use mobile phone to access information related to farming? 1=Yes, 2=No If NO, go to C10 C9. a. If YES, what type of information? (Tick all that apply) 1=Weather, 2=Price, 3=Inputs, 4=Production technology, 5=Pest control, 6=Government programs, 99= Others (Specify) 4 times, 3=5-6 times, 4=7-10 times, 5=More than 10 times C10. How many plots/parcels of land does the Household own (including rented / borrowed/shared/leased in) C9. b. From whom do you access information using mobile phone(Tick all that apply) 1=Relatives/friends, 2=Input dealers, 3=KVK, 4=Kisan Call Center, 5=Extension agents, 6=RML, 7=IFFCO Kisan Sanchar Limited (IKSL), 8=mKRISHI, 9=Other mobile based agro advisory services, 99=Others (Specify) C11. On how many plots / parcels of land has the farmer cultivated groundnuts in Kharif 2015 season? Particulars for each plot on which groundnut is cultivated in Kharif 2015 (total Number of columns should equal the response in C12) C12. Total Area (size of this plot) (in local units) C13. Area under Groundnut Cultivation C14. Distance from your home to this plot (in kms) 68 Plot # 1 Plot # 2 Plot # 3 Plot # 4 Plot # 5 C15. Land Tenure 1=Owned, 2=Crop sharing (yearly basis), 3=Leased for cash (yearly basis), 4=Pledged without interest, 99=Others (Specify) C16. Is groundnut inter/mixed/border crop cultivated with other crops on this plot? (1=Yes, 2=No) if YES, go to next question, if NO go to C18: C17. Name of the inter/mixed Crop with Groundnut 1=Red gram, 2=Green gram, 3=Horse gram, 4=Cow pea, 5= Black gram, 6= Bengal gram, 7= Finger millet, 8= Little millet, 9= Foxtail millet, 10= Pearl millet, 11= Barnyard millet, 12=Kodo millet, 13= Proso millet, 14=Sorghum,15=Maize, 16=Not cultivated any inter/mixed crop, 99=Others (Specify) C18. Border Crop 1=Red gram, 2=Green gram, 3=Horse gram, 4=Cow pea, 5= Black gram, 6= Bengal gram, 7= Finger millet, 8= Little millet, 9= Foxtail millet, 10= Pearl millet, 11= Barnyard millet, 12=Kodo millet, 13= Proso millet, 14=Sorghum,15=Maize, 99=Others (Specify) C19. Name of the Groundnut Variety 1=Anantha Jyothi (ICGV 91114),2=Bheema (TG 47), 3=Dharani,4=JL 24, 5=Kadiri 3, 6=Kadiri 5,7=Kadiri 6, 8=Kadiri 7,9=Kadiri 8, 10=Kadiri 9,11=Local (nondescriptive),12=Narayani, 13=Pollachi red (local land race),14=TGCS 1043, 15=TMV 2, 99=Others (Specify) C20. Source of groundnut seed planted 1=Saved from previous harvest, 2=Purchased from market as oilseed, 3=Purchased from other farmers or community based organization who produced seed, 4=Purchased from seed companies or input dealers, 5=Received subsidized seed from government or NGOs, 6=Purchased from Agriculture Office, 7=Borrowed / obtained from neighbors/relatives, 8=Do not remember, 99=Others (specify) C21. Can you recall when was the first year you adopted this variety? 9999=Do not remember C22. What was the source of seed for this variety when you first adopted it 1=Purchased from market as oilseed, 2=Purchased from other farmers or community based organization who produced seed, 3=Purchased from seed companies or input dealers, 4=Received subsidized seed from government or NGOs, , 5=Purchased from Agriculture Office, 6=Borrowed / obtained from neighbors/relatives,7=Do not remember, 99=Others (Specify) C23. What are the two characteristics of this variety you LIKE? 1=High yield, 2= Resistance to Insect and disease, 3=Drought resistance, 4=Early maturity, 5=Seed quality, 6=Color and taste, 7=Processing quality, 8=Good price / high demand, 99=Others (Specify) C24. What are the two characteristics of this variety you DISLIKE? 1=Low yield, 2=Susceptible to insects and diseases, 3=Susceptible to drought, 4=Late maturity, 5=Seed quality, 6=Color and taste, 7=Processing quality, 8=Low price / Low demand, 99=Others (Specify) C23.a. First C23.b. Second C24. a. First C24.b.Se cond C25. Source / type of Irrigation 1=Rainfed, 2=Bore well, 3=Well, 4=Canal, 5= River, 6=Stream, 7=Lake/ponds, 8=Pit,9=Drip irrigation, 10=Sprinkler, 99=Others (Specify) C26. Who in your house mainly C26.a. Persons relationship to the main decision maker 1=Self, 2=Spouse, 3=Son/Daughter, 4=Brother/Sister, 5=Parent, 99=Others (Specify) C26.b. Gender (1=Male, 2-Female) 69 provides labour for this plot? C27. Who is the main decision maker regarding inputs and outputs of this plot C27.a. Persons relationship to the main decision maker 1=Self, 2=Spouse, 3=Son/Daughter, 4=Brother/Sister, 5=Parent, 99=Others (Specify) C27.b. Gender (1=Male, 2-Female) C28.a. Organic fertilizer C28.b. Chemical fertilizer C28. Have you used or plan to use any of these C28.c. Pesticides inputs on this plot this Kharif season (Check all that C28.d. Herbicides apply) C28.e. Hired labour C28.f. Other inputs (specify) C29. Expected groundnut production on this plot (in kgs) C30. In your assessment is your soil type fit for cultivating Groundnut? 1=Yes, 2= No C31. How would you rate the soil quality of this plot:1=Low, 2=Medium, 3=High, 98=Don’t know SECTION D: TECHNOLOGY ADOPTION D1. Are you currently using the following technologies on this plot? (Check all that apply) 1=Soil bunds, 2=Field/boundary bunds, 3=Broad bed and furrow, 4=Land leveling, 5=Contour bunds, 6=Polythene mulching, 7=Nala plugs/RFDs, 8=Sunken pits, 9=Farm pond, 10=Masonry check dams, 11=Well recharge pits, 12=Penning Sheep/Goat/Cattle, 13=Others (Specify)…………………… TX.1. In your estimate what is the total area on your farm covered by this technology? 1=Less than 20%, 2= 20-40%, 3=40-60%, 4=60-80%, 5=More than 80% TX.2. First year of adoption of _________ on your farm (9999=Do not remember) TX.3. What is the main source of information about the technology of ____________? 1=Agricultural office, 2=Farmer Cooperative/Union, 3=Farmer group/association, 4=Learnt from Training programme, 5=Learnt from demonstration, 6=Traditionally known, 7=NGO/CBO, 8=Another farmer relatives and friends, 9=Another farmer neighbor, 10=Radio/newspaper/TV, 99= Others (Specify) TX.4. In your assessment what are the main benefits of using ____________ on your farm 1=Saves water, 2=Soil management, 3=Improves crop yield, 4=More uniform moisture-environment for crops, 5=Reduces weeds problems, 6=Easy land preparation, 7=Improves uniformity of crop growth and maturity, 8=Reduces labour cost,9=Reduces cost of other inputs, 99= Others (Specify) TX.5. In your opinion does the ____________ increase, decrease or has no effect on the labour time devoted to farming by the MALE members of your household? 1=Increase, 2=Decrease, 3=Neutral TX.6 If response is increase or decrease, ask: In what aspects is the labour input by male members increased or decreased? 1=Land preparation, 2=Sowing, 3=Weeding, 4= Irrigation, 5=Harvesting 70 TX.7. In your opinion does the ____________ increase, decrease or has no effect on the labour time devoted to farming by the FEMALE members of your household? 1=Increase, 2=Decrease, 3= Neutral TX.8. If response is increase or decrease, ask: In what aspects is the labour input by female members increased or decreased? 1=Land preparation, 2=Sowing, 3=Weeding, 4= Irrigation, 5=Harvesting D2. Have you ever used the following technologies in the past? (Check all that apply) 1=Soil bunds, 2=Field/boundary bunds, 3=Broad bed and furrow, 4=Land leveling, 5=Contour bunds, 6=Polythene mulching, 7=Nala plugs/RFDs, 8=Sunken pits, 9=Farm pond, 10=Masonry check dams, 11=Well recharge pits, 12=Penning Sheep/Goat/Cattle, 13=Others (Specify)…………………… TX.9. When was the last year you used ________ (9999=Do not remember) TX.10. Reason for dis-adopting of __________ technology 1=Low yield, 2=High cost involved, 3=Insects affect, 4=More wastage of net area sown, 5=Not accessible to Machineries, 6=Not aware of this, 99=Others (Specify) D3. Other than the technologies mentioned above, has your household adopted any NEW practices, inputs, farming methods promoted by the agricultural extension service, KVKs, a research center, or a private input dealer? 1=Yes 2=No If YES go to next question; if NO, go to next section D3.1. If yes, what was this new technology? D3.2. When did you first adopt / use this technology on your farm? (year) SECTION E: CONNECTIONS E1. How many farmers in this village/other villages you know personally? E1.a. How many days once you are interacting with them on farming related issues? 1=Regularly, 2=Twice in a week, 3=3 times in a week, 4= 4 times in a week, 5= 5 or more than 5 times in a week, 99=Do not interact E2. Totally how many farmers you know have used/currently using the following technologies on their farm? (read each, and note the Numbers) E2.a. Soil bunds E2.b. Field/boundary bunds E2.c. Broad bed and furrow E2.d. Land leveling E2.e. Contour bunds E2.f. Polythene mulching E2.g. Nala plugs/RFDs E2.h. Sunken pits E2.i. Farm pond E2.j. Masonry check dams E2.k. Well recharge pits E2.l. Penning Sheep/Goat/Cattle E2.j.Others (Specify)…………………… SECTION F: MIGRATION F1. Has any member of your household migrated in the past 12 months? section G 71 (1=Yes, 2=No) If YES go to (next question), if NO go to F2. What was the primary reason for migration? 1=To earn wages as food grain, 2=To earn higher wages, 3=To reduce burden on family, 4=Non-availability of work in village, 5=To work on relative’s farm, 6=Other (specify) SECTION G: POVERTY SCORE CARD G1. How many people aged 0-17 are present in your household? 0=Five or more, 4=Four, 8=Three, 13=Two, 20=One, 27=None G2. What is household’s principal occupation? 0=Labourers (agricultural plantation, others farm), hunters, tobacco preparers and tobacco product makers and other labourers, 14=Professionals, technicians, clerks, administrators, managers, executives, directors, supervisors and teachers 8=Others G3. Is the residence all pucca (burnt bricks, Stone, Cement, Concrete, Jack board/Cement-plastered reeds, timber, tiles, gal vanished tin or asbestos cement sheets)? 4= Yes, 0=No G4. What is the household’s primary source of energy for cooking? 0= Firewood and chips, Charcoal or none, 17=LPG, 5=Others G5. Does the household have own television? 6= Yes, 0=No G6. Does the household own a bicycle, scooter or motor cycle? 5= Yes, 0=No G7. Does the household own an almirah/dressing table? 3= Yes, 0=No G8. Does the household own a sewing (tailoring) machine? 6= Yes, 0=No G9. How many pressure cookers or pressure pans does the household own 0=None, 6=One, 9=Two or more G10. How many electric fans does the household own? 0=None, 5= One, 9= Two or more SECTION H: OTHER RELATED QUESTIONS H1. During the past Kharif 2014 season what was the total groundnut produced by your household? (in Bags) H2. How much of this was sold? (in Bags) H3. What are the two most important constraints you face in groundnut farming? 1=Land, 2=Labour, 3=Cash constraint, 4=Seeds not available, 5=Insect and disease problem, 6=Cannot sell the crop, 7=Price is too low, 8=No information or technical advice on farming practices, 9=Low rainfall, 10=Not accessible to the machinery, 99=Others (Specify) H4. Have you stopped growing ANY groundnut varieties in the past 3 years that you used to grow before? 1= Yes, 2= No; if yes go to next question; if no go to H7 H4.a. If YES, how many? H5. Give me the name of recent variety you have discontinued and the MAIN reason for not growing it anymore? 1=Anantha Jyothi (ICGV 91114),2=Bheema (TG 47), 3=Dharani,4=JL 24, 5=Kadiri 3, 6=Kadiri 5,7=Kadiri 6, 8=Kadiri 7,9=Kadiri 8, 10=Kadiri 9,11=Local (non-descriptive),12=Narayani, 13=Pollachi red (local land race),14=TGCS 1043, 15=TMV 2, 99=Others (Specify) H6. Reason codes: 1=Seed not available, 2=Had low yield, 3=Did not like the color, 4=Susceptible to diseases, 5=Not liked by processors, 6=Unpleasing cooking quality/taste, 7=Not adopting to climatic conditions, 99=Others (specify) G5.a. Name G6.a. Reason H7. What is the distance from your house to the nearest paved road? (if the house is next to the paved road, write zero) (in kms) H8. What is the distance from your house to the nearest market where you obtain agricultural inputs (e.g., fertilizer, pesticides, seeds, etc.)? (in kms) H9. What is the distance from your house to the nearest agricultural extension office? (in kms) H10. What means of transport do you mainly use to get to the nearest commercial town? 1=Walking, 2=Tractor, 3=Bicycle, 4=Motorcycle, 5=Car, 6=Bus, 7=Light transport Vehicle, 8=Animal Driven Cart, 99=Others (specify) H11. Distance from your home to this commercial town? (in kms) 72 H12. Time it takes on average to travel to this commercial town using the main mode of transportation H12.a. Hours H12.b. Minutes H13. Indicate if this survey was completed in the first attempt or required a re-visit 1= First attempt, 2= Re-visit H14. Indicate if this HH was part of the first randomly selected farmer or a replacement 1= First random selection, 2= Replacement THANK YOU VERY MUCH FOR YOUR TIME End time of interview 73 APPENDIX B: CEA Implementer – Groundnuts Questionnaire To be filled by enumerator Date of the interview (dd/mm/yyyy) Name of the enumerator Time started To be filled by Supervisor Date checked (dd/mm/yyyy) Name of the supervisor Section 1: Current household composition and characteristics 1.1 Household identification 1a District CODE A 1b. Mandal / Block 1c. Gram Panchayat 2b. CODE 2a Village/hamlet name 3 Indicate random selection of Household Elevation (in meter) GPS coordinate N (Format xx.xxxxx) CODE B Household (HH) id District Code 4 4 6 GPS coordinate E (Format xx.xxxxx) Mandal / Block Village Code Code HH Number Code A: 1-Karnal, 2-Ludhiana, 3-Vaishali, 4-Kurnool, 5-Anantapur Code B: 1-Original, 2-Replacement 1.2 General information about the Respondent Note: Respondent here refers to the Lead Decision Maker for Agricultural Activities in the family. 1 Name of the Respondent a. First Name b. Last Name 2 3 4a 4b 4c 5 6 7a 7b 8 9a 9b 9c 10 11 Gender of the Respondent (Code: 1 –Male, 2- Female) Age of the Respondent Marital status (CODE A) How many brothers and sisters do you have (including siblings that may a. brothers b. sisters have died)? What was your birth order? For example, were you first born, second born, third…? Years of formal education of Respondent and spouse (if a. Respondent b. Spouse (if married) married) Main occupation (CODE B) a. Respondent b. Spouse (if married) Can read a local Indian language (1-Yes, 2-No) Can read English (1-Yes, 2-No) Years of Experience in farming Years of experience in growing groundnut Years of experience in growing wheat Years of experience in growing rice Mobile number Relationship with Head of the household (HOH) (CODE B) If option 1, skip to section 17 74 12 13 14 15 Name of the Head of Household Gender of the Head of the HH (Code: 1 –Male, 2- Female) Age of the Household Head Years of formal education 16a 16b 17a a. Head b. Spouse of Head (if married) Can read a local Indian language (1-Yes, 2-No) Can read English (1-Yes, 2-No) Religion of the household (Code: 1- Hindu, 2- Muslim, 3-Christian, 4- Sikh, 5-Buddhist 98-Don’t know, 99Others (specify) 17b Caste (Code 1- General ,2- SC ,3- ST, 98-Don’t Know,99 –Others) 18 Highest level of formal education completed by any a. b. Gender (1-Male 2member of the household (Years of education), and the Education Female) gender of that individual 19 Are you a member of any farmer organization or a farmer cooperative (1-Yes 2-No) 20 If Yes, what is your level of involvement in this group? 1- very active, 2-somewhat active, 3-not active 21 Are you a leader of any of these groups? (1-Yes 2-No) 22 How many sons and daughters do you have a. Sons b. Daughters 23 If farmer has both sons and daughters ask: What is the birth order of your eldest son? Is he 1 st born, 2nd living born,…? Code A: 1-Married with spouse, 2-Married but spouse away, 3-Divorced, 4-Widow, 5-Not married, 99-other (specify), Code B:1- Farming on own farm, 2- Livestock rearing, 3- Salaried employment, 4- Self-employed off farm, 5- Casual labourer on-farm, 6- Casual labour off farm, 99-other (specify). Code C: 1-Head himself/herself, 2- Wife, 3- Husband, 4- Son, 5-Daughter, 6- Grandchild, 7- Father, 8-Mother, 9- Sister, 10Brother, 11- Niece, 12- Nephew, 13- Son in law,14- Daughter in law, 15-Brother in law, 16-Sister in Law, 17- Father in law, 18- Mother in law, 19- Other family relatives, 20- Servant, 21- Permanent labour, 22-Tenants, 99- Other person not related 1.3. Household Information A 'household' is usually a group of (related or unrelated) persons who normally live together and take their meals from a common kitchen. If a group of unrelated persons live in a census house but do not take their meals from the common kitchen, then they are not constituent of a common household. 1.3.1 How many members belong to this household (first write total number, then by age and gender groups): ____ Total By age and gender FEMALE: a. <5 years ____ b. 5-17 ______ c. >18 _____ d. Total female members_____ MALE: e. <5 years ____ f. 5-17 ______ g. >18 _____ h. Total male members_____ 1.3.2 Total working members in the family (in Numbers) a. Male________ b. Female _________ 1.3.3 In the past 12 months, did any member of your household obtain income from any of the following sources? (Instruction: Read each item and note yes/no) 1=Yes 0=No a. Sale proceeds of Field Crops b. Horticulture crop sales 1=Yes g. Wages from off farm (govt. job, teacher, etc) h. Non-farm business or self-employment i. Remittance j. Pension Income k. Other (specify) c. Dairy d. Livestock sales for meat e. Renting/leasing land or farm equipment f. Wages from farm labor 1.3.4 Total annual household income across all the activities and working members (Rs) a. Cash: (CODE A) b. in kind (cash equivalent) (CODE A) 75 0=No Code A: 0 = < 25,000 1= 25,000-50,000, 2=50,000-1,00,000, 3= 1,00,000- 2,00,000, 4= 2,00,000-3,00,000, 5= 3,00,000-4,00,000, 6= 4,00,000-5,00,000, 7= 5,00,000 – 6,00,000, 8= 6,00,000 – 8,00,000 , 9= 8,00,000 – 10,00,000, 10= greater than 10,00,000 1.3.5. What source of income mentioned above contributes the largest share to your total household income? (write a code a to k corresponding to the source mentioned) ________ 1.3.6 In your estimate, what percentage of your total HH income in the last 2 years came from? a. groundnut farming ___________ 1.3.6 What type of Ration card do you have? __________ 1-APL (white), 2-BPL (blue), 3-AYY (yellow), 4-AY (special), 5-Do not have ration card, 99-Others (Specify) 1.3.7 Is anyone in your household (other than you) a member of a farmer producer organization or a farm cooperative?__________ 1-Yes, 2-No 1.4 Poverty Score Card (The codes correspond to the poverty SCORE, PLEASE KEEP THESE SCORE CODES when programming the survey) 1 2 3 4 5 6 7 8 9 10 How many people aged 0-17 are currently part of your household? 0-Five or more, 4-Four, 8-Three, 13-Two, 20-One, 27-None What is household’s principal occupation? 0-Laborers (agricultural plantation, others farm), hunters, tobacco preparers and tobacco product makers and other labourers 14-Professionals, technicians, clerks, administrators, managers, executives, directors, supervisors and teachers 8-Others Is the residence all pucca (burnt bricks,Stone, Cement, Concrete, Jack board/Cement-plastered reeds, timber, tiles, gal vanished tin or asbestos cement sheets)? 4-Yes, 0-No What is the household’s primary source of energy for cooking? 0-Firewood and chips, Charcoal or none, 17-LPG, 5-Others Does the household have own television? 6-Yes, 0-No Does the household own a bicycle, scooter or motor cycle? 5-Yes, 0-No Does the household own an almirah/dressing table? 3-Yes, 0-No Does the household own a sewing (tailoring) machine? 6-Yes, 0-No How many pressure cookers or pressure pans does the household own 0-None, 6-One, 9-Two or more How many electric fans does the household own? 0-None, 5-One, 9-Two or more 1.5. Information on Migrant Family Member (Enumerator: A member is usually termed as migrated if she/he lives outside village for more than a year or left recently with that intention) Only for related family members .Exclude spouses /children of migrant members) 1a 1b Has any member of your household migrated in the past 5 years? (1Has any member of your household migrated in the past 12 months? Yes, 2-No) If NO to both these questions, skip to next section (2.1) (1-Yes, 2-No) 2 Place of most recent migration CODE A 3 Reasons for migration CODE B 4 Does the member who has migrated take major decisions in matter relating to the agricultural activities? (1-Yes, 2-No) 5 Does the member who has migrated contribute towards meeting household Code A: 1-Within state (urban 2- Within state (rural area), 3-Within country, another state, 4-Middle East, 5expenses? (1-Yes,area), 2-No) US/Canada/Australia, 6-European Countries, 99. Others (specify) Code B: 1- Better prospect of employment, 2- Weather related uncertainties, 3- Higher education, 4- Marriage, 98- Don’t know, 99-Others (sp) 76 Section II. Land Holding 2.1 Land ownership 2.1a 2.1b 2.1c 2.1d 2.1e 2.1g 2.1f Local land unit (LU) CODE A LU conversion rate How many Plots of land does your household own? How much area of land your household owns across all these plots?much (LU) land that your household currently owns was: How Acquired through purchase? Inherited? Acquired through other means (specify)? 1 acre =...................LU Total across 2.1e to 2.1g should equal 2.1d Have you ever sold any land that you had inherited? If No, write 0, If Yes, indicate the total land area sold ; 999-did not inherityou anyever landsold any land that you had acquired through purchase? If No, 2.1i Have write 0, If Yes, indicate the total land area sold; 999-have not purchased any 2.1j If YES to either 2.1h or 2.1i: What was the main reason for selling the land? land CODE B 2- Acre, 3- Killa, 4- Kanal, 5-Bissa, 99- Others (specify) Code A: 1- Bigha, 2.1h Code B: 1-to pay off debt; 2- to get cash for non-farm business or investment; 3-to meet household expenses; 4-to downsize my farming operation; 5-Other (specify) III. Technology specific questions 3.4. Technology X11 S. Questions N 1 2 3 4 5 7 8a 8b 9 10 10 11 b 12 12 24 b 27 Have you ever heard about TECHNOLOGY X? (1-Yes, 2-N0) If NO, skip to next When did you first come to know about it? (9999-Don’t Know) YYYY section Source of information CODE A Have you ever used TECHNOLOGY X? (1-Yes, 0- No) If YES, skip to 7; if NO, ask 5 and skip to next technology If No, why? CODE B (main reason) When did you start using it? (9999-Don’t Know) YYYY Who was involved in making the decision to use TECHNOLOGY X? CODE C What was the motivation behind the decision to use TECHNOLOGY X? CODE D Did you stop using it once you adopted it?(1-yes, 0-No) if NO, skip to 11 If yes, why? (main reason) CODE B When was the last year you used TECHNOLOGY X? (9999=Do not remember) Have you or anyone in your household received training in using TECHNOLOGY X?Yes, (1-Yes, 0-No) If NO, skip to If From whom? (CODE A, 13 1 to 8) Are you currently using TECHNOLOGY X on your farm? (1=Yes, 2= No) If NO, go to 27 In your estimate what is the total area on your farm covered by this technology? 1=Less than 5%, 2= 5-10%, 3=10-15%, 4=15-20%, 5=20-40%, 6-40-60%, 7-60-90%, 890-100%, 99-not applicable In your opinion, what are the main 6-Increase nutrient efficiency benefits of using TECHNOLOGY X? 1-Better Uniformity of crop growth / maturity 11 7-Reduces weed problem 8-Labor saving 9-Easy land preparation This question box was repeated for each of the 12 technologies in the original questionnaire 77 2-Reduce water requirement / saves water and cost 3-Improves crop establishment 4-Increases water application efficiency 5-Higher Yields 28 29 30 31 31 a 31 b 31 c 31 d 31 e 32 f 10-reduces cost of other inputs 99-Others (Specify) 97-No more benefits Do you face any inconveniences in using SOIL BUND technology? (Rank Top 3) 1-Unavailable at the peak time 2-too expensive 3-Service provider does not provide credit 4-Lack of service provider/materials in the village 5-Unsatisfied with technology 6-Unsatisfied with service quality 7-Difficulty in getting subsidy 8-lack of repair and service facility nearby 9-Frequent technical problems 10-Labor intensive 99-Others (Specify) 97-No inconvenience Do you share your SOIL BUND technology experience with other farmers? (1- Yes, 0-No) In your opinion, does the SOIL BUND technology increase, decrease or has no effect on the time devoted to farming by the MALE members of your household relative to the conventional practice? (Code 1-Increase, 2-Decrease, 3-Neutral)If Neutral, skip to 32 If response is increase or decrease: In what aspects land is the labor input by MALE members increased or Sowing preparation decreased? Weeding 1-Yes, 0-No Irrigating Harvesting Other (specify) In your opinion, does the SOIL BUND technology increase, decrease or has no effect on the time devoted to farming by the FEMALE members of your household relative to the conventional practice? (Code: 1-Increase, 2-Decrease,3-Neutral)If Neutral, skip to next section If response is increase or decrease: In what aspects land is the labor input by FEMALE members increased or preparation Sowing decreased? Weeding 1-Yes, 0-No Irrigating Harvesting Other (specify) Rank Top 3 33 33 a 33 b 33 c 33 d 33 e Code f A: 1-Government Extension service, 2- Service Provider, 3- CIMMYT/ICRISAT, 4- Farmer Cooperative /group, 5-Research centres other than CIMMYT/ICRISAT, 6- Neighbour/Relative farmer, 7- Private Company/input dealer, 8- NGO/CBO, 9- Radio, 10- TV, 11- Mobile Phone , 12- Newspaper,, 13-traditionally known, 99- Others (Specify) Code B: 1-Unwilling to try new technology, 2-lack training/information, 3- Expensive to hire/build, 4- Service/materials not available in the village, 5-Gives Less Yield, 6- Not satisfied with output, 7- Does not look good, 8- High weed, 9- Not suitable on small Land, 10- Not suitable for the crop, 11- Does not have irrigation facility, 12- Land is naturally level/ no need, 13-Difficulty in getting subsidy, 14- Lack of information 98-Cannot say, 99- Others (Specify) Code C: 1-Only I myself made the decision, 2-Both me and my spouse were involved, 3-I and other male members of my family made the decision, 4-Whole family was involved, 99-Others (Specify) Code D: 1-to increase crop yield/productivity, 2-to reduce irrigation cost or water wastage, 3-to control weed problem, 4other farmers in the village were using it, 5-Other (specify) 78 IV. Plot characteristics and groundnut production in Kharif 2015 season 4.1 Land use 1a How many plots did you cultivate in Rabi 2014-15 season? 1b What was the total cultivated land area in Rabi 2014-15 season (LU) 2a How many plots did you cultivate in Kharif 2015 season? 2b What was the total cultivated land area in kharif 2015 season (LU) 3 Did you leave any land fallow in Rabi 2014 and this Kharif 2015 season? 1-Yes 2-No (If NO, go to 4.2) 4 Reason for leaving land fallow (1- Land not fertile, 2- Unavailability of water, 3Dispute over land, 4-Unabvailability of labour, 99-Others (Specify) What crops did your HH produce in the last 12 months in the following categories (only record number of crops mentioned): 5a. Cereal 5b. Pulses 5c. oil 5d. horticulture 5e. fibre crops 5f. Other crops 4.2. Plot characteristics For each of the plot cultivated in Rabi 2014-15 and Kharif 2015 season, I would like to ask you some specific questions. S.N 1a 1b 2 3a 3b 3c 4 5 6 7 8 9 10 11 12 13a 13b Note to Enumerators: Start with biggest plot farmer cultivated Distance from your home to this plot km Size /Area of plot (LU) Plot ownership CODE: 1- Owned, 2.-Leased in /Shared in, 3- Leased out/Shared Out by CODE A If owned, Owned If owned, was this plot inherited or purchased? 1-inherited 2purchased 98-don’t know Your assessment of the market value of this plot if you were to sell it today? Rs Irrigation source CODE: 0- No irrigation, 1-Tube Well, 2-Open Well, 3River canal water, 4-Pond 99- Others (specify) Irrigation type CODE: 1-Flood, 2-Pump, 3-Drip, 4-Sprinkler, 99-Other (specify) Soil type CODE: 1-Sandy, 2-Sandy Loam, 3-Loam Soils, 4-Clay Loam, 5Clay, 99-Others (Specify) Soil quality CODE:1- Good, 2-Medium, 3-Poor Soil Salinity CODE:1- High, 2- Medium, 3- Low 98-don’t know Land level CODE:1- High level, 2- Middle level, 3- Low level, 4uneven/mixed, 5-uniform 98-don’t know What is your observation about the soil quality, fertility on this plot compared to last 10 years? (Code 1. Declining, 2. Remain same, 3. Improving) How many times this plot has been leveled? If ZERO, ask 12 then go to 14a ; other than ZERO, skip to 13a Why this plot has never been leveled? CODE B (Multiple responses possible) When was the plot leveled last? (mm/yyyy) Method of leveling used CODE:(0 –Traditional, 1-Laser land levelling, 2-None) 79 Plot Plot Plot I 2 3 S.N 14a 14b 14c 14d Note to Enumerators: Start with biggest plot farmer cultivated Plot Plot Plot I 2 3 What did you do with the crop residues on this plot at the end of Kharif 2014? CODES: 0-No residue was produced, 1-Retained/incorporated in the field, 2-Mulched, 3-Burnt it, 4-Used it as fodder or cooking fuel onfarm, 5-Sold it, 99-Other (specify) (multiple responses are possible) If retained, incorporated or mulched: Percentage retained/mulched/incorporated? What did you do with the crop residues on this plot at the end of Rabi 2015? CODES: 0-No residue was produced, 1-Retained in the field, 2Mulched, 3-Burnt it, 4-Used it as fodder or cooking fuel on-farm, 5Sold it, 99-Other (specify) If retained or mulched: Percentage retained/mulched? 15a Soil bunds, 15b Field/boundary bunds 15c Broad bed and furrow 15e Are you currently using Contour bunds 15f following technologies on this Polythene mulching 15g plot or more generally on your Nala plugs/RFDs 15h farm with direct impact on this Sunken pits plot? (Check all that apply) 115i Farm pond Yes, 2-No 15j Masonry check dams 15k Well recharge pits 15l Penning Sheep/Goat/Cattle 15m Others (Specify)…………………… 16a Do you have following types of Fruit trees 16b planted trees on this plot? Trees for firewood/fuel 16c Trees for soil fertility 16d CODE: 1-Yes 2-No Trees for commercial purpose Code A: 1-Head himself/herself, 2- Wife, 3- Husband, 4- Son, 5-Daughter, 6- Grandchild, 7- Father, 8-Mother, 9- Sister, 10Brother, 11- Niece, 12- Nephew, 13- Son in law, 14- Daughter in law, 15-Brother in law, 16-Sister in Law, 17- Father in law, 18- Mother in law Code B: 0-Don’t know what is Laser levelling / this service is not available here; 1- Financial constraint, 2- Used on trial basis, 3-Land is naturally level, 4- Not required for particular crop, 5-Small land size, 6- Land not empty/vacant, 7-leased land , 99-Others(Specify) Following questions are specific to groundnut farming in Kharif 2015(continue with the same plots) S.N Note to Enumerators: Start with biggest plot farmer cultivated Plot I Plot 2 96 97 98 99 99 a Is this plot cultivated with groundnut in Kharif 2015? CODE: 1-Yes 2-No If No, skip to Next Plot or Section Is groundnut inter-cropped? CODE: 1-Yes 2-No If NO, skip to 99 What percentage of this plot is planted to groundnut in the Kharif season? Name of the inter/mixed Crop CODE C 98=Not cultivated any inter/mixed Border crop crop CODE C 98=Not cultivated any border crop b 80 Plot 3 S.N Note to Enumerators: Start with biggest plot farmer cultivated 10 Did you practice crop rotation on this plot by planting a legume crop before or after a cereal crop? CODE: 1-Yes 2-No 0 10 1 10 2 10 10 3 4 10 5 Method of Plot Preparation in Kharif 2015 CODE: 1-Planker 2- Tiller/Cultivator, 3- Rotavator, 4- Harrow, 5Paddy Harrow, 6-ZT drill 7-LLL 8-conventional ploughing 9-ripping 10-ridging 99- others (Specify) Record multiple response upto 4 forCODE:1-Broadcasting, each plot Method of seeding in Kharif 2015 2-seed cum ferti drill,3-ZT drill, 4-Turbo happy seeder Seeding rate (kg/LU) Who in your house mainly provided labour for this plot in Kahrif 2015? (write the relationship with the respondent and gender) a. Relationship (CODE A) Who is the main decision maker regarding inputs and outputs of this plot? (write the relationship with the respondent and gender) a. Relationship (CODE A) 10 6 10 7 10 b. Gender (1-Male, 2Female) b. Gender (1-Male, 2Female) What groundnut variety of seed is planted in Kharif 2015? CODE D 8 10 9 11 0 11 1 11 2 Source of groundnut seed planted CODE: 1=Saved from previous harvest, 2=Purchased from grain vendors in the market, 3=Purchased from other farmers or community based organization who produced seed, 4=Purchased from seed companies or input dealers, 5=Received subsidized seed from government or NGOs, 6=Received seed from extension agents, 7=Borrowed / obtained from neighbors/relatives, 98=Do not remember, 99=Others (specify) Can you recall when was the first year you adopted this variety on your farm? 9998= less than 15 years ago but do not remember; 9999= more than 15 years ago but do not remember; What was the source of groundnut seed for the first planting? CODE: 1=Purchased from grain vendors in the market, 2=Purchased from other farmers or community based organization who produced seed, 3=Purchased from seed companies or input dealers, 4=Received subsidized seed from government or NGOs, 5=Received seed from extension agents, 6=Borrowed / obtained from neighbors/relatives, 98=Do not remember, 99=Others (specify) What are the two characteristics of this variety you First LIKE? 81 Plot I Plot 2 Plot 3 S.N Note to Enumerators: Start with biggest plot farmer cultivated 11 CODE: 1=High yielding, 2= Resistance to Insect and Second disease, 3=Drought resistance, 4=Early maturity, 5=Seed quality, 6=Colour and taste, 7=Processing quality, 8=Good price / high demand, 99=Others (Specify) What are the two characteristics of this variety you First DISLIKE?CODE: 1=Low yielding, 2=Susceptible to insects and diseases, 3=Susceptible to drought, 4=Late maturity, 5=Seed quality, 6=Color and taste, Second 7=Processing quality, 8=Low price / Low demand, 99=Others (Specify) Did you use any of a=organic fertilizer (kg) these inputs / b=Urea (Kg) practices on this plot c=DAP (kg) in Kharif 2015 (read each d=Potash (Kg) input and note the response e=Phosphate (Kg) 0-No, if Yes, note the total Quantity of input used) f= Zinc (kg) 3 11 4 11 5 11 11 6a 11 6b 11 6c 11 6d 11 6e 11 6f 11 6g 11 6h 11 6i 11 6j 11 6k 116 6l 11 m 11 6n 11 6o 11 7a 11 7b 11 7c 11 7d 11 7e 11 7f 11 8 9a 11 12 9b 0 g=pesticides h=herbicides i=hired labor (yes/no) j=Nutrient Expert Decision Support software k=Leaf Colour Chart (yes/no) l=GreenSeeker sensors (yes/no) i=hired labor (yes/no) What was the Total Man Days of land preparation Labor required in the following Sowing activities for groundnut farming Weeding in Kharif Season 2015? Irrigating Harvesting 96-Not Applicable Other (specify) How many times was this plot irrigated in Kharif Season 2015? What is the total quantity of groundnut you expect to harvest from this plot in this season? Qtl Specify if the weight is with or without shell: 1=with shell, If plot is inter-mixed crop, ask: What is the total value of other 2=without shell crops you expect to harvest from this plot in this season? Rs Plot I Plot 2 Plot 3 (choose unit mg/gm/ml/lit/kg) (choose unit mg/gm/ml/lit/kg) Code A: 1-Head himself/herself, 2- Wife, 3- Husband, 4- Son, 5-Daughter, 6- Grandchild, 7- Father, 8-Mother, 9- Sister, 10Brother, 11- Niece, 12- Nephew, 13- Son in law, 14- Daughter in law, 15-Brother in law, 16-Sister in Law, 17- Father in law, 18- Mother in law Code C: 1=Red gram, 2=Green gram, 3=Horse gram, 4=Cow pea, 5= Black gram, 6= Bengal gram, 7= Lentils, 8=kidney beans / Rajma, 9=pigeon pea; 9= groundnut, 10=mustard, 11= Finger millet, 12= Little millet, 13= Foxtail millet, 14= Pearl millet, 15= Barnyard millet, 16=Kodo millet, 17= Proso millet, 18=Sorghum,19=Maize, 20-wheat, 21-rice, 22- soybean, 98=Not cultivated any inter/mixed/border crop, 99=Others (Specify) Code D: 1=Anantha Jyothi (ICGV 91114),2=Bheema (TG 47), 3=Dharani,4=JL 24, 5=Kadiri 3, 6=Kadiri 5,7=Kadiri 6, 8=Kadiri 7,9=Kadiri 8, 10=Kadiri 9,11=Local (non-descriptive),12=Narayani, 13=Pollachi red (local land race),14=TGCS 1043, 15=TMV 2, 99=Others (Specify) 82 Section V: Perception on new technologies, constraints and access to information and credit 5.1. Perception of New Agriculture Practices S.N Questions .1. When was the last time your Household adopted a NEW input or a farming COD E practice on your farm for the first time? (YYYY) 2. What was this new input or farming practice you most recently adopted on your farm? CODE: 1-seed/Variety, 2-Agro-chemicals, 3-New animal breed, 4agronomic practices, 5-soil or water conservation, 6-conservation agriculture, 7-Machinery/tools, 8-Storage method, 9-mono-cropping, 10drying/processing, 99-Others (Specify) 3 What is the depth of ground water level in this area?(ft) 98 Don’t know 4 Over the past 10 years have you experienced fall in ground water level?(2No, 98-Don’t Know, If YES, by how much ft water level has declined) If NO, skip to 6 5 If YES, In your opinion, what are the reasons for the decline in ground water level? CODE A of any water conservation practices? (1- Yes, 0-No) if NO, 6 Are you aware skip toname next section 7 If yes, the practices? (Record up to three) (CODE A B C B) you using any of these practices on your farm? CODE: 1-Yes 2-No 8 Are 9a If YES, which one(s)? 9b If NO, why not? CODE C 10 Have you ever used hybrid seeds of any crop on your farm? CODE: 1-Yes 2Code A: 1-No Indiscriminate use of water, 2. Deforestation, 3- Increase in population, 4-Increased industrial activity, 5- Water Pollution, 6-Decline in rainfall, 7-Reason not related to human activity,8- Increase in submersible pump, 98-Don’t Know, 99Others (specify) Code B: 1-Scheduling irrigation only when required, 2. Planting less water requiring crops/variety, 3. Keeping residue for water conservation, 4- Adopting water saving technology, 5- Farm pond, 99. Others (specify) Code C: 1- Beyond my control, 2- Single effort will not help, 3- Water saving technology are costly, 4- My land uses surface water, 5- No water problem in my area, 99- Others (specify) 5.2 Use of technology by farmers in social network 1. How many farmers in this village/other villages you know personally? 2. How often you are interacting with them on farming related issues? (1=at most once a week, 2=Twice in a week, 3=3 times in a week, 4= 4 times in a week, 5= 5/more than 5 times in a week, 6=Do not interact) 3. Approximately how many farmers YOU KNOW have used/currently using the following technologies on their farm? (read each, and note the numbers) Code: 997--I am not aware of this practice myself, 999—I am aware about this practice but don’t know how many are using it a=Zero Tillage i Soil bunds b1=Land levelling j. Field/boundary bunds b2=Laser land levelling k. Broad bed and furrow c=Direct seeding rice l. Contour bunds d=Residue retention/mulching m. Polythene mulching e=legume rotation n. Nala plugs/RFDs f=drip irrigation o. Sunken pits g=green seeker p. Farm pond h=leaf color chart q. Masonry check dams i=Nutrient expert decision support software r. Well recharge pits j=agroforestry s. Penning Sheep/Goat/Cattle h=hybrid seeds 83 5.3c Constraints to Groundnut production S.No 1a 1b 2 3 4 5 6a 6b 6c Questions What are the two main constraints you face in groundnut First farming? CODE A Second Have you stopped planting any groundnut varieties in the past 3 years that you used to grow before?(1-Yes,0-No) If NO, skip to 5.5 If yes, How many? Name the most recent variety you have discontinued What is the main reason for discontinuation? CODE B What is the current price of groundnut (with shell) if you were to sell it? Rs/kg What is the current price of groundnut (with shell) if you were to buy it? Rs/kg What is the selling price of groundnut (with shell) you expect at the time of harvest this season? Rs/kg Code A: 1-Land, 2-labour, 3-cash constraint, 4-seeds not available, 5-insect and disease problem, 6-cannot sell the crop, 7-price too low, 8-no information or technical advice on farming practices, 99-other(specify) Code B: 1-Seed not available, 2-had low yield, 3-did not like the color, 4-susceptible to disease, 5- not liked by processors, 6-unpleasing cooking quality/taste, 99-other (specify) 5.3d Sources of risk in farming and coping strategies 1a 1b 1c 1d 1e 1f 1g 1h 1i 1j 2a 2b 2c 2d 2e 2f 3 4 Variability in the timing and level of rainfall Floods Drought High temperatures Hail or cold temperatures Insects and plant diseases Infectious livestock diseases Price fluctuations in farm commodities of inputs in a Non-availability Lack market to sell the products timelyofmanner When you face an Sell household goods, jewellery, etc. economic shock due to Sell animals/ livestock any of these risk factors Sell other farm assets mentioned above, Borrow money what coping strategies I change my farming practices by going back to doing things the do you most often use? traditional way I change my practices by using NEW and MODERN methods of (indicate 1=Use most farming If use 2e often or sometime, ask: Can you give an example of this strategy you have used in the often, 2=use sometime, 3-Never) past? If use 2f often or sometime, ask: Can you give an example of this strategy you have used in the past? Based on your experience, which of the following events would you consider to be a major cause of concern to you as a farmer or a major source of risk in your farming operation? Read each event and ask the respondent to rank them on a scale of 0 to 2 0 – not a cause of concern for me 1- Somewhat a concern for me 2- A major concern for me 5.4 Loss Due to Unexpected Weather 1. 2. 3. In the past 5 years, have you ever lost a significant portion of your crop production due to unexpected weather (e.g., low rainfall, flooding, unexpected monsoon time, hail, etc)? 1-Yes, only wheat, rice or groundnut, 2-Yes, Other crop, 3- Yes, multiple crops, including wheat, rice, groundnut, 4-No If yes, how many times you have suffered such losses in past five years? 98-Don’t know/can’t say/don’t remember How many times you have suffered such losses in the past two years? 98-Don’t know/can’t say/don’t remember 84 4 Which crops were most impacted by these losses? 1-wheat, 2-rice, 3-groundnut, 4other, 5-all 5a. Have you heard about the phenomenon called ‘climate change’? 1-Yes 2-No 5b If YES, can you tell me what will happen as a result of ‘climate change’ that has 5c implications for farmers like you? (select as many as mentioned): 5d 1-extreme weather; 2-too much rain/flood; 3-too little rain/drought; 4-high 5e temperatures; 5-late start of rainy season; 6-cold winters; 7-too much pest/diseases; 8-unpredictable weather; 9-farming will become more risky; 10-Other (specify); 98Don’t know 5.5 Access to information, infrastructure and credit 1 What is the distance from your house to the nearest paved road (if the house is next to the paved 2 3 4 5 6a 6b 7 8 9 10 11 12 13 14 15 a 15 b 15 c 15 d 16 a road, write zero) km What is the distance from your house to the nearest market where you obtain agricultural inputs (e.g., fertilizer, pesticides, seeds, etc.) km What is the distance from your house to the nearest agricultural extension office km What means of transport do you mainly use to get to the nearest commercial town? 1- Walking, 2- Tractor, 3- Bicycle, 4-Motorcycle, 5-Car, 6- Bus, 7- Light transport Vehicle, 8-Animal Driven Cart, 99- others (specify) Distance from your home to this commercial town km Time it takes on average to travel to this commercial town using the main mode of Hours transportation (Consider time of one way travel) minutes Do you have a bank account? Code: 1-Yes 2-No Do you own kisan credit card? (1-Yes, 2-No) Do you currently have crop insurance policy (other than KCC)? Code: 1-Yes 2-No, If YES, skip to 12 Did you have crop insurance in the past but have discontinued? Code: 1-Yes 2-No Reason for not having crop insurance or for discontinuing:1=Do not need, 2=Not available in my place, 3= No claim available at time of damage 4=Too expensive, 99= Others (Specify) Did you or anyone in the household access credit for agricultural production in the past 12 months (1=Yes, 2= No) If NO, skip to 14 If yes, from where? 1- Bank, 2- Cooperatives, 3- SHG, 4-Community member, 5-Relative/ Friend/ Neighbour, 6-Local money Lender, 7-Commission Agent, 8- Employer, 9-Agrovet, 10-Trader 99- Others (Specify) If no, why not? 1= Did not need, 2= Do not have access to credit, 3=Very high interest rate , 4-Far From Residence, 5Bank staff not cooperative, 6- was getting less amount 7-loan was not approved 8-no collateral 99= Others (Specify) Do you currently have any debt (i.e., do you owe money to anyone)? 1-Yes 2-No If NO, skip to 17a If yes, to whom do you owe money? 1- Bank, 2- Cooperatives, 3- SHG, 4-Community member, 5-Relative/ Friend/ Neighbour, 6-Local money Lender, 7-Commission Agent, 8- Employer, 9-Agrovet, 10-Trader 96-multiple (specify) 99Others (Specify) What is the interest PER ANNUM you are paying on this debt? In how many years do you expect to pay-off this debt? If you need to borrow money for any purpose, how likely is it that you will be able to borrow money you need? (Read the possible responses and select one) 1-Extremely likely (about 100% chance), 2-Quite likely (about 75% chance), 3-Neither likely nor unlikely (about 50%), 4-Quite unlikely (about 25% chance), 5-Extremely unlikely (about 0%) (skip to 17a) 85 16 b 17 a 17 b 19 20 21 22 23 Who will be the main source of credit? 1- Bank, 2- Cooperatives, 3- SHG, 4-Community member, 5Relative/ Friend/ Neighbour, 6-Local money Lender, 7-Commission Agent, 8- Employer, 9-Agrovet, 10-Trader 99- Others (Specify) During the past year, where did you receive most of your information and advice about groundnut or general agricultural production and marketing from? CODE A (Multiple Response) In total, approximately how many times did you receive information about agricultural production and marketing last year from all these sources? Do you use mobile phone to access information related to farming? 1=Yes 2=no If NO, go to 21 If yes, what type of information? CODE C – Multiple Response From whom do you access information using mobile phone? CODE D— Multiple Response Have you heard about the following organizations / programs? (1-Yes 2-No ) a. CIMMYT b. ICRISAT c. IRRI d. CGIAR e. CCAFS f. Climate Smart Villages g. Krishi Vignan Kendra h. Internet What is the farthest you have ever travelled? 0-never left this village; 1-a village/town in this district; 2-a village/town in a neighboring district; 3-a neighbouring state; 4-another state within India; 5-another country in South Asia; 6-US/Canada/Australia/Europe; 7-Middle east; 9-Other (specify) 24 What is the farthest anyone who is currently a member of your household (other than you) has travelled so far and his/her relationship to you? 0-never left this village; 1-a village/town in this district; 2-a village/town in a neighbouring district; 3-a neighbouring state; 4-another state within India; 5-another country in South Asia; 6US/Canada/Australia/Europe; 7-Middle east; 9-Other (specify) 25 a.travel b. relationship (CODE E) When it comes to adopting new technology, inputs or farming practices, which of the following best describes your behaviour: 1 - I am one of the first ones to adopt NEW technologies 2 – I usually wait until a few farmers I know have used those inputs/technologies/practices, and then based on their experiences I make the decision 3 – I usually wait until most farmers in this village are already using those inputs/technologies/practices, and I am 100% sure that those technologies work 4 – I rarely change my farming practices as I am not comfortable doing new things 26 Do you have life insurance policy? 1-Yes, 2-No CODE A. 1-Extension agent, 2-NGO staff, 3-Trader / input dealer, 4-Farmer group/leader farmer, 5-Service Provider, 6-I did not receive any information or advice, 99-other (specify) CODE B. 1=1-2 times, 2=2-3 times, 3=3-5 times, 4=5-10 times, 5=More than 10 times CODE C. 1-Weather, 2-Price, 3-Inputs, 4-Production technology, 5-Pest control, 6-government programs, 99other (Specify) CODE D. 1-Relatives/friends, 2-Input dealers, 3-KVK, 4-Kisan Call Center, 5-Extension agents, 6-RML, 7-IFFCO Kisan Sanchar Limited (IKSL), 8-mKRISHI, 9-other mobile based agro advisory services, 99-Other (specify) 86 Code A: 2-wife, 3-Husband, 4- Son, 5-Daughter, 6- Grandchild, 7- Father, 8-Mother, 9- Sister, 10-Brother, 11- Niece, 12- Nephew, 13- Son in law, 14- Daughter in law, 15-Brother in law, 16Sister in Law, 17- Father in law, 18- Mother in law 5.6 Assets owned How many of the following does your household own and what is the total value (in Rs) if you were to sell it today? (Instruction: For each item, write the number owned and its total value across all units owned. If none owned, write zero) 1 2 3 4 5 6 7 Bicycle Motor cycle / Scooter Car / truck Cart Tractor Plough Metal silos Water tank Irrigation / water pump 8 9 10 Greenhous Dehusker e / glass house 11 12 Fodder chopper 13 14 15 16 17 18 19 20 21 22 23 24 Combine harvester Cultivator / Zero till tiller drills Biogas plant Turbo/ Happy seeder Seed-cum- LLL Ferti Drills Ripper Radio / cassette player TV Fans AC 25 26 27 28 29 30 31 32 33 34 35 36 Cooler Washing machine Water purifier Camera Pressure cooker Almirah Refrigerator Comp-uter Sewing machine Gas stove Mobile phones NonMobile phone # owned value # owned value # owned value 87 How many animals does your household currently own and its total value across all units owned (if none owned, write zero) 36 38 39 40 41 42 43 44 45 Horses Cows Buffaloes Bulls Goats Sheep # owned value Donkeys / Mules To be filled by Enumerator after the completion of Survey End Time Was the survey completed in first attempt or required a revisit? 1-First Attempt, 2-ReVisit THANK YOU FOR PARTICIPATING IN THIS SURVEY 88 Pigs Chickens 46 Other (describe) APPENDIX C: LEA Implementer 2 Questionnaire To be filled by enumerator Date of the interview (dd/mm/yyyy) Name of the enumerator dd mm yyyy Time started Enumerator’s acquaintance with farmer CODE A Code A: 1= not met before, 2=yes-in the context of another survey, 3=yes-in a social context, 4=yes-we are personal acquaintances To be filled by Supervisor Date checked (dd/mm/yyyy) Name of the supervisor Section 1: Current household composition and characteristics 1.1 Household identification 1 District 2 Block name 3 Village name 4 GPS coordinate N (Format xx.xxxxx) 5 Household (HH) id CODE A District Code CODE CODE GPS coordinate E (Format xx.xxxxx) Block Code Village Code HH Number Code A: 1-Karnal, 2-Ludhiana 1.2 General information about the Head of the Household Note: Respondent here refers to the Lead Decision Maker for Agricultural Activities in the family. 1 Name of the Respondent First Name Last Name 2 3 4 5 Gender of the Respondent (Code: 1 –Male, 2- Female) Age of the Respondent Marital status (CODE A) Years of formal education a. Respondent b. Spouse (if married) 6 Relationship with HOH (CODE B) If option 1, skip to 11 Name of the Head of Household Gender of the Head of the HH (Code: 1 –Male, 2- Female) Age of the Household Head Years of formal education a. Head b. Spouse (if married) Religion of the household (Code: 1- Hindu, 2- Muslim, 3-Christian, 4- Sikh, 5Buddhist 98-Don’t know, 99- Others (specify) Main occupation (CODE C) a. Respondent b. Spouse (if married) 7 8 9 10 11 12 89 13 14 15 16 Years of Experience in farming Years of experience in growing wheat Years of experience in growing rice Mobile number Code A: 1-Married living with spouse, 2-Married but spouse away, 3-Divorced, 4-Widow, 5-Not married, 99-other (specify), Code B: 1-Head himself/herself, 2- Wife, 3- Husband, 4- Son, 5-Daughter, 6- Grandchild, 7- Father, 8-Mother, 9- Sister, 10Brother, 11- Niece, 12- Nephew, 13- Son in law,14- Daughter in law, 15-Brother in law, 16-Sister in Law, 17- Father in law, 18- Mother in law, 19- Other family relatives, 20- Servant, 21- Permanent labour, 22-Tenants, 99- Other person not related Code C:1- Farming on own farm, 2- Livestock rearing, 3- Salaried employment, 4- Self-employed off farm, 5- Casual labourer on-farm, 6- Casual labour off farm, 7- Involved in household chores, 99-other (specify). General Remarks (anything that is noteworthy such as cropping pattern, no of plot and their respective size, if land is taken in lease if yes area of leased land, family size etc) 1.3. Household Information A 'household' is usually a group of (related or unrelated) persons who normally live together and take their meals from a common kitchen. If a group of unrelated persons live a census house but do not take their meals from the common kitchen, then they are not constituent of a common household. 1.3.1 How many members belong to this household ________ 1.3.2 Total working members in the family (in Numbers) a. Male________ b. Female _________ 1.3.3 Total annual household income across all the activities and working members (Rs) a. Cash: (CODE A) b. in kind (cash equivalent) (CODE A) Code A: 1= 0, 2= 1-50,000, 3=50,000-1,00,000, 4= 1,00,000- 2,00,000, 5= 2,00,000-3,00,000, 6= 3,00,000-4,00,000, 7= 4,00,000-5,00,000, 8= greater than 5,00,000 1.3.4 In the past 12 months, did any member of your household obtain income from any of the following sources? (Instruction: Read each item and note yes/no) 1=Yes 0=No a. Sale proceeds of Field Crops b. Horticulture crop sales c. Dairy d. Livestock sales for meat e. Renting/leasing land f. Renting/leasing farm equipment 1=Ye s g. Wages from farm labor h. Wages from off farm (govt. job, teacher,etc) i. Non-farm business or selfemployment j. Remittance k. Pension Income l. Other (specify) 1.3.5 In your estimate, what percentage of your total HH income in the last 2 years came from? a. wheat farming ________ b. rice farming _________ 90 0=No 1.3.6 What type of Ration card do you have? __________ 1-APL (GREEN), 2-BPL(YELLOW), 3-AYY(Pink), 4- AY (SPECIAL) 5- OPH(KHAKI), 6-Do not have Ration card, 99-OTHERS(SPECIFY) 1.3.7 Does anyone in your household is a member of a farmer producer organization or a farm cooperative?__________ 1-Yes, 0-No 1.4 Poverty Score Card (I have changed the codes to correspond to the SCORE, PLEASE KEEP THESE SCORE CODES when programming the survey) 1 2 3 4 5 6 7 8 9 10 How many people aged 0-17 are currently part of your household? 0-Five or more, 4-Four, 8-Three, 13-Two, 20-One, 27-None What is household’s principal occupation? 0-Laborers (agricultural plantation, others farm), hunters, tobacco preparers and tobacco product makers and other labourers 14-Professionals, technicians, clerks, administrators, managers, executives, directors, supervisors and teachers 8-Others Is the residence all pucca (burnt bricks,Stone, Cement, Concrete, Jack board/Cementplastered reeds, timber, tiles, gal vanished tin or asbestos cement sheets)? 4-Yes, 0-No What is the household’s primary source of energy for cooking? 0-Firewood and chips, Charcoal or none, 17-LPG, 5-Others Does the household have own television? 6-Yes, 0-No Does the household own a bicycle, scooter or motor cycle? 5-Yes, 0-No Does the household own an almirah/dressing table? 3-Yes, 0-No Does the household own a sewing (tailoring) machine? 6-Yes, 0-No How many pressure cookers or pressure pans does the household own 0-None, 6-One, 9Two or more How many electric fans does the household own? 0-None, 5-One, 9-Two or more 1.5. Information on Migrant Family Member (Enumerator: A member is usually termed as migrated if she/he lives outside village for more than a year or left recently with that intention) Only for related family members .Exclude spouses /children of migrant members) 1a Has any member of your household migrated in the past 5 years? (1Yes, 0-No) 1b Has any member of your household migrated in the past 12 months? (1-Yes, 0-No) If NO to both these questions, skip to next section (2.1) 2 Place of most recent migration CODE A 3 Reasons for migration CODE B 4 Does the member take major decisions in matter relating to the agricultural activities? (1-Yes, 0-No) Code A: 1-Within state (urban area), 2- Within state (rural area), 3-Within country, 4-Middle East, 5-US/Canada/Australia, 6European Countries, 99. Others (specify) Code B: 1- Better prospect of employment, 2- Weather related uncertainties, 3- Higher education, 4- Marriage, 98- Don’t know, 99-Others (specify) 91 Section II. Land Holding 2.1 Land ownership 2.1a Local land unit (LU) CODE A 2.1b LU conversion rate 2.1c How many Plots of land does your household own? 2.1d How much area of land your household owns across all these plots? (LU) Code A: 1Bigha, 2- Acre, 3- Killa, 4- Kanal, 5-Bissa, 99- Others (specify) 1 acre =...................LU III. Technology specific questions 3.1. Technology X12 S.N 1 2 3 4 5 6 7 8 9 Have you ever heard about TECHNOLOGY X? (1-Yes, 0-N0) If NO, skip to next When did you come to know about it? (9999-Don’t Know) section (3.2) Source of information CODE A Have you ever used TECHNOLOGY X? (1-Yes, 0- No) If YES, skip to 7; if NO, ask 5 and 6 and skip to nextreason) section (3.2) If No, why? CODE B (main Will you adopt it if there is access to a service provider? (1-Yes, 0- No, 98-Don’t know/can’t say) When did you start using it? (9999-Don’t Know) Who was involved in making the decision to use TECHNOLOGY X? CODE C Did you stop using it once you adopted it?(1-yes, 0-No) if NO, skip to 11 10 If yes, why? (main reason) CODE B 11 Have you or anyone in your household received training in using TECHNOLOGY X? (1-Yes, 0-No) If NO, skip to 13 If Yes, From whom? (CODE A, 1 to 7,99-Others(Specify)) 12 13 YYYY YYYY 15b When was the last time you used TECHNOLOGY a. Month b. Year X (indicate month and year) Did you use your own leveler, borrowed or hire it? (code 1-own 2-Hired 3-Borrowed) If own, skip to 17; If hired, ask 15 and then skip to 24; if Borrowed, ask 16 and then skip to 24 What was the per unit cost of hiring Select the Unit TECHNOLOGY X when you leveled last? 1=hour 2=acre 99=other (specify) Rupees per unit 15c From whom did you hire the TECHNOLOGY X? CODE D 16 From whom did you Borrow the TECHNOLOGY X? CODE D 17 At what price did you purchase the TECHNOLOGY X Machine? (Rs) 18 When did you purchase it? (YYYY) 19 Did you receive any subsidy at the time of purchase?(0-No, if YES amount of subsidy) Do you lease this machine to others on rental basis? 1-Yes 0-No, If NO, skip to 24 14 15a 20 12 Questions This question box was repeated for each of the 3 technologies in the original questionnaire 92 21 22a 22b 23 24a 24b When was the last time you rented the TECHNOLOGY X to others? (MM/YYYY) What was is the per unit revenue you Select the Unit earn from renting out the TECHNOLOGY 1=hour 2=acre 99=other X when you rented to others last time? (specify) Rupees per unit you charged For how many units did you rent out your TECHNOLOGY X last time you rented to others? Who mainly operates TECHNOLOGY X on Age your farm? Capture age and gender Gender 1=Male 2=Female 25 On how many plots on your farm did you use this technology when last time you used it? 26 What is the total area cultivated using this technology when last time you used it? LU The last time you used TECHNOLOGY X, which crop was planted? CODE a b E 27 28 In your opinion, what are the main benefits of using laser land leveler? (Rank Top 3) Suggested Benefits Rank Top 3 Uniformity of growth Reduce water requirement Higher Yields Same or more output with lesser inputs Reduce weed Labor saving Others (Specify) No Response 29 Do you face any inconveniences in using TECHNOLOGY X? (0-No, If YES, Rank Top 3) Suggested Inconveniences Unavailable at the peak time too expensive to hire Service provider does not provide credit Lack of service provider in the village Unsatisfied with technology More time than required taken Service provider not levelling the field properly Difficulty in getting subsidy lack of repair and service facility nearby Frequent technical problems with Machine Others (Specify) 93 Rank Top 3 c d No Response 30 31 32a 32b 32c Do you share your TECHNOLOGY X experience with other farmers? (1- Yes, 0-No) In your opinion, does the laser land leveler increase, decrease or has no effect on the time devoted to farming by the MALE members of your household relative to the conventional practice? (Code 1-Increase, 2-Decrease, 3-Neutral)If Neutral, skip to 33 If response is increase or decrease: In what land aspects is the labor input by MALE members preparation increased or decreased? Sowing 1-Yes, 0-No Weeding 32d Irrigating 32e Harvesting 32f Other (specify) 33 34a 34b 34c In your opinion, does the laser land leveller increase, decrease or has no effect on the time devoted to farming by the FEMALE members of your household relative to the conventional practice? (Code: 1-Increase, 2-Decrease,3-Neutral)If Neutral, skip to Section 35 If response is increase or decrease: In what land aspects is the labor input by FEMALE members preparation increased or decreased? Sowing 1-Yes, 0-No Weeding 34d Irrigating 34e Harvesting 34f Other (specify) 35 In your opinion, what is the impact of this technology on labor? (Code: 1-Increase, 2-Decrease,3-Neutral) Code A: 1-Government Extension service, 2- Service Provider, 3- CIMMYT, 4- Farmer Cooperative /group, 5-Research centres other than CIMMYT, 6- Neighbour/Relative farmer, 7- Private Company, 8- Radio, 9- TV, 10- Mobile Phone , 11- Newspaper, 99- Others (Specify) Code B: 1-Unwilling to try new technology, 2- Expensive to hire, 3- Not available in the village, 4-Gives Less Yield, 5- Lack of service provider, 6- Not satisfied with output, 7- Does not look good, 8- High weed, 9- Not suitable on small Land, 10- Not suitable for the crop, 11- Does not have irrigation facility, 12- Land is naturally level, 13-Difficulty in getting subsidy, 14- Access to information 98-Cannot say, 99- Others (Specify) Code C: 1-Only I myself made the decision, 2-Both me and my spouse were involved, 3-me and the Head of Household, 4-I and other male members of my family made the decision, 5-Whole family was involved, 99-Others (Specify) Code D: 1-Service provider in village, 2- Service provider from other village, 3-Village cooperative, 4-Relatives/Neighbour farmer, 5-Farmers association, 6-Progressive farmer, 7-Farmer Cooperative, 8-Government Extension service, 99-Others (Specify) Code E: 1-Rice, 2-Wheat, 3-Pulses, 4-Vegetables, 5-Fodder, 96-Not Applicable, 99-Others (Specify) 94 IV. Plot characteristics and wheat/rice production in Rabi 2014-15 and Kharif 2015 season 4.1 Land use 1a How many plots did you cultivate in Rabi 2014-15 season? 1b What was the total cultivated land area in Rabi 2014-15 season (LU) 2a How many plots did you cultivate in Kharif 2015 season? 2b What was the total cultivated land area in kharif 2015 season (LU) 3 What crops your HH produced and harvested in the last 12 months in the following categories (only record number of crops mentioned) Categories No. of Crops Categories a. Cereal crops No. of Crops b. Pulse crops d. Horticulture crops (incl, veg, fruits, herbs) e. Fibre crops c. Oil seed crops f. Other 4.2. Plot characteristics For each of the plot cultivated in Rabi 2014-15 and Kharif 2015 season, I would like to ask you some specific questions. S.N 1 2 3 4 5 6 7 8 9 10 11 12a Note to Enumerators: Start with biggest plot farmer cultivated Size /Area of plot (LU) Plot ownership CODE: 1- Owned, 2.-Leased in /Shared in, 3Leased out/Shared Owned/Leased by Out CODE A Soil type CODE: 1-Sandy, 2-Sandy Loam, 3-Loam Soils, 4-Clay Loam, 5- Clay, 99-Others (Specify) Soil quality CODE:1- Good, 2-Medium, 3-Poor Soil Salinity CODE:1- High, 2- Medium, 3- Low, 4-No Soil Salanity 98-don’t know Land level CODE:1- High level, 2- Middle level, 3- Low level 98don’t know What is your observation about the soil quality, fertility on this plot compared to last 10 years? (Code 1. Declining, 2. Remain same, 3. Improving) How many times this plot has been laser leveled? If ZERO, ask 10 and skip to 12; other than ZERO, skip to 11 Why this plot has never been laser leveled? CODE B (Multiple responses possible) When was the plot leveled last? (mm/yyyy) What did you do with the crop residues on this plot at the end of Kharif 2014? CODES: 0-No residue was produced, 1-Retained in the field, 2Incorporated 3-Mulched, 4-Burnt it, 5-Used it as fodder or cooking fuel on-farm, 6-Sold it, 7-Select All, 99-Other (specify) (multiple responses are possible) 95 Plot Plot Plot I 2 3 S.N Note to Enumerators: Start with biggest plot farmer cultivated Plot Plot Plot I 2 3 12b If retained, mulched or incorporated: Percentage retained/mulched/incorporated? CODES: 1- 0-25%, 2- 25-50%, 350-75%, 4- 75-100% 13a What did you do with the crop residues on this plot at the end of Rabi 2014-2015? CODES: 0-No residue was produced, 1-Retained in the field, 2Incorporated 3-Mulched, 4-Burnt it, 5-Used it as fodder or cooking fuel on-farm, 6-Sold it, 7-Select All, 99-Other (specify) (multiple responses are possible) 13b If retained, mulched or incorporated: Percentage retained/mulched/incorporated? CODES: 1- 0-25%, 2- 25-50%, 350-75%, 4- 75-100% 14 Do you have following types of planted trees on this field? CODE: 1-Yes 0-No 14a Fruit trees 14b Trees for firewood/fuel 14c Trees for soil fertility 14d Trees for commercial purpose Code A: 1-Head himself/herself, 2- Wife, 3- Husband, 4- Son, 5-Daughter, 6- Grandchild, 7- Father, 8-Mother, 9- Sister, 10Brother, 11- Niece, 12- Nephew, 13- Son in law, 14- Daughter in law, 15-Brother in law, 16-Sister in Law, 17- Father in law, 18- Mother in law Code B:1- Financial constraint, 2- Used on trial basis, 3-Land is naturally level, 4- Not required for particular crop, 5-Small land size, 6- Land not empty/vacant, 7-leased land , 99-Others(Specify) Following questions are specific to wheat farming in Rabi 2014-2015 (continue for the same plots) S.N Note to Enumerators: Start with biggest plot farmer Plot I Plot 2 Plot cultivated 3 15 Was this plot cultivated with wheat in Rabi 2014-15? CODE: 1-Yes 0-No If No, Skip to 56 16 If yes, what was the area under Wheat Cultivation? (LU) 17 Was wheat inter-cropped? CODE: 1-Yes 0-No If NO, skip towhat 20 was the total value of other crops 18 If YES, harvested on this plot in Rabi 2014-2015? 9819 What percentage of this plot was planted to wheat in don’t remember the season?crop rotation on this plot by planting 20 Did you practice Rs Rabi a legume crop before or after wheat crop? CODE: 1Yes 0-No 21 Irrigation source CODE: 0- No irrigation, 1-Tube Well, 2-Open Well, 3- River canal water, 4-Pond 99- Others (specify) 22 Irrigation type CODE: 1-Flood (with pump), 2-Flood (without pump), 3-Furrow, 4-Drip, 5-Sprinkler, 99-Other (specify) 23 Method of Plot Preparation in Rabi 2014-2015 96 24 25 26 27 28 29 CODE: 1-Planker 2- Tiller/Cultivator, 3- Rotavator, 4Harrow, 5-Paddy Harrow, 6-ZT drill 7-LLL 8-TLL 9conventional ploughing 10-Ripper 11-Bund Maker 99- others (Specify) Record multiple response up to 5 for each plot Method of seeding in Rabi 2014-2015 CODE:1Broadcasting,2-seed cum ferti drill,3-ZT drill, 4-Turbo happy seeder, 99-others Seeding rate (kg/LU) Who in your house mainly provided labour for this plot? (write the relationship Head of the Household) CODE A Who is the main decision maker regarding inputs and outputs of this plot? (write the relationship Head of the Household) What wheat variety of seed was planted in Rabi 20142015? Source of wheat seed planted CODE: 1=Saved from previous harvest, 2=Purchased from market, 3=Purchased from other farmers or community based organization who produced seed, 4=Purchased from seed companies or input dealers, 5=Received subsidized seed from government or NGOs, 6=Received seed from extension agents, 7=Borrowed / obtained from neighbors/relatives, 98=Do not remember, 99=Others (specify) 30 31 32a 32b 33a 33b Can you recall when was the first year you adopted this variety on your farm?9999=Do not remember What was the source of wheat seed for the first planting? CODE: 1=Purchased from market, 2=Purchased from other farmers or community based organization who produced seed, 3=Purchased from seed companies or input dealers, 4=Received subsidized seed from government or NGOs, 5=Received seed from extension agents, 6=Borrowed / obtained from neighbors/relatives, 98=Do not remember, 99=Others (specify) What are the two characteristics of First this variety you LIKE? Second CODE: 1=High yielding, 2= Resistance to Insect and disease, 3=Drought resistance, 4=Early maturity, 5=Seed quality, 6=Colour and taste, 7=Processing quality, 8=Good price / high demand, 100-No response, 99=Others (Specify) What are the two characteristics of First this variety you DISLIKE?CODE: 1=Low Second yielding, 2=Susceptible to insects and diseases, 3=Susceptible to drought, 4=Late maturity, 5=Seed quality, 6=Color and taste, 7=Processing quality, 97 8=Low price / Low demand, 100-No response, 99=Others (Specify) S.No Questions For 34-41 record two responses for each plot. One response should be for Per LU and the Second response for Total Quantity 34 Did you use any of a=organic fertilizer (kg) these inputs / 35 b=Urea (Kg) practices on this 36 c=DAP (kg) plot in Rabi 37 d=Potash (Kg) season (read each input 38 e=Phosphate (Kg) and note the response) 39 f= Zinc (kg) 0-No, if Yes, note the per LU/total Quantity of input 40 used) Plot 1 Plot 2 Plot 3 Per Ttl Per Ttl Per Ttl LU Qty LU Qty LU Qty Name of Pesticide (choose unit 1-mg/2-gm / 3-ml / 4-lit/5-kg) g=pesticides(record the name and quantity of the top 3 pesticides used for the crop) Name of Pesticide (choose unit 1-mg/2-gm / 3-ml / 4-lit/5-kg) Name of Pesticide (choose unit 1-mg/2-gm / 3-ml / 4-lit/5-kg) 41 Name of Herbicide (choose unit 1-mg/2-gm / 3-ml / 4-lit/5-kg) Name of Herbicide h=herbicides(record the name (choose unit 1-mg/2-gm / 3-ml / and quantity of the top 3 4-lit/5-kg) herbicides used for the crop) Name of Herbicide (choose unit 1-mg/2-gm / 3-ml / 4-lit/5-kg) 42 43 44 45 46 47 i=Nutrient Expert Decision Support software j=Leaf Colour Chart k=GreenSeeker sensors l=other inputs (specify) What was the Total Man Days of Labor land required in the following activities for preparation wheat farming in Rabi Season 2014-15? Sowing 48 Weeding 49 Irrigating 98 50 51 harvesting 54 Other (specify) How many times was this plot irrigated in Rabi Season 2014-15? What was the total quantity of wheat produced from this plot in Rabi 2014-15? Qtl Wheat quantity sold Qtl 55 Price sold Rs/Qtl 52 53 Code A: 1-Head himself/herself, 2- Wife, 3- Husband, 4- Son, 5-Daughter, 6- Grandchild, 7- Father, 8-Mother, 9- Sister, 10-Brother, 11- Niece, 12- Nephew, 13- Son in law, 14- Daughter in law, 15-Brother in law, 16-Sister in Law, 17- Father in law, 18- Mother in law Code B:1- Financial constraint, 2- Used on trial basis, 3-Land is naturally level, 4- Not required for particular crop, 5Small land size, 6- Land not empty/Vacant, 99-Others(Specify) Following questions are specific to rice farming in Kharif 2015(continue with the same plots) S.N Note to Enumerators: Start with biggest plot farmer cultivated Plot Plot Plot I 2 3 56 Is this plot cultivated with rice in Kharif 2015? CODE: 1-Yes 0-No If No, skip to Next section (5.1) 57 If yes, what was the area under Rice Cultivation? (LU) 58 Is rice inter-cropped? CODE: 1-Yes 0-No If NO, skip to 60 59 What percentage of this plot is planted to rice in the Kharif season? 60 Did you practice crop rotation on this plot by planting a legume crop before or after the rice crop? CODE: 1-Yes 0-No Irrigation source CODE: 0- No irrigation, 1-Tube Well, 2-Open Well, 3River canal water, 4-Pond 99- Others (specify) Irrigation type CODE: 1-Flood (with pump), 2-Flood (without pump), 3Furrow, 4-Drip, 5-Sprinkler, 99-Other (specify) Method of Plot Preparation in Kharif 2015 CODE: 1-Planker 2- Tiller/Cultivator, 3- Rotavator, 4- Harrow, 5-Paddy Harrow, 6-ZT drill 7-LLL 8-TLL 9-conventional ploughing 10-ripper 11-Bund Maker 99- others (Specify) Record multiple response upto 5 for each plot Method of seeding in Kharif 2015 CODE:1-Broadcasting,2-seed cum ferti drill,3-ZT drill, 4-Turbo happy seeder 99-Other (specify) Seeding rate (kg/LU) 61 62 63 64 65 66 68 Who in your house mainly provided labour for this plot? (write the relationship with the Head of the Household) CODE A Who is the main decision maker regarding inputs and outputs of this plot? (write the relationship with the name and relationship with the ) CODE A What rice variety of seed is planted in Kharif 2015? 69 Source of rice seed planted 67 CODE: 1=Saved from previous harvest, 2=Purchased from market, 3=Purchased from other farmers or community based organization who produced seed, 4=Purchased from seed companies or input dealers, 5=Received subsidized seed from government or NGOs, 6=Received seed from extension agents, 99 S.N Note to Enumerators: Start with biggest plot farmer cultivated Plot I Plot 2 Plot 3 7=Borrowed / obtained from neighbors/relatives, 98=Do not remember, 99=Others (specify) 70 Can you recall when was the first year you adopted this variety on your farm? 9999=Do not remember 71 What was the source of rice seed for the first planning? CODE: 1=Purchased from market, 2=Purchased from other farmers or community based organization who produced seed, 3=Purchased from seed companies or input dealers, 4=Received subsidized seed from government or NGOs, 5=Received seed from extension agents, 6=Borrowed / obtained from neighbors/relatives, 98=Do not remember, 99=Others (specify) 72a What are the two characteristics of this variety you First LIKE? 72b Second CODE: 1=High yielding, 2= Resistance to Insect and disease, 3=Drought resistance, 4=Early maturity, 5=Seed quality, 6=Colour and taste, 7=Processing quality, 8=Good price / high demand, 100-No response, 99=Others (Specify) 73a What are the two characteristics of this variety you First DISLIKE?CODE: 1=Low yielding, 2=Susceptible to insects 73b Second and diseases, 3=Susceptible to drought, 4=Late maturity, 5=Seed quality, 6=Color and taste, 7=Processing quality, 8=Low price / Low demand, 100-No response, 99=Others (Specify) S.No Questions For 74-81 record two responses for each plot. One response should be for Per LU and the SecondDid response for Total Quantity 74 you use a=organic fertilizer (kg) any of 75 b=Urea (Kg) these 76 c=DAP (kg) inputs / 77 d=Potash (Kg) practices on 78 e=Phosphate (Kg) this plot in 79 f= Zinc (kg) Kharif 2015 80 (read each input and note the response 0-No, if Yes, note the total Quantity of input used) 81 g=pesticides (record the name and quantity of the top 3 pesticides used for the crop) Plot 1 Plot 2 Plot 3 Per Ttl Qty Per LU Ttl Per Ttl LU Qty LU Qty Name of Pesticide (choose unit 1-mg/2-gm / 3-ml / 4-lit/5kg) Name of Pesticide (choose unit 1-mg/2-gm / 3-ml / 4-lit/5kg) Name of Pesticide (choose unit 1-mg/2-gm / 3-ml / 4-lit/5- kg) Name of Herbicide h=herbicides(record the name (choose unit 1-mg/2-gm / 3-ml / 4-lit/5and quantity of the top 3 herbicides used for the crop) kg) Name of Herbicide 100 (choose unit 1-mg/2-gm / 3-ml / 4-lit/5kg) Name of Herbicide (choose unit 1-mg/2-gm / 3-ml / 4-lit/5kg) 82 83 84 85 86 87 88 89 90 91 92 93 94 What was the Total Man Days of Labor required in the following activities for wheat farming in Kharif Season 15? i=Nutrient Expert Decision Support software j=Leaf Colour Chart k=GreenSeeker sensors l=other inputs (specify) land preparation Sowing Weeding Irrigating Harvesting Other (specify) 96-Not Applicable How many times was this plot irrigated in Kharif Season 2015? What is the total quantity of rice you expect to harvest from this plot in this season? Qtl Specify if the weight is husked rice or dehusked: 1=husked 2=dehusked Section V: Perception on new technologies, constraints and access to information and credit 5.1. Perception of New Agriculture Practices S.N. Questions 1. When was the last time your Household adopted a NEW input or a farming practice on your farm for the first time? (YYYY) 2. What was this new input or farming practice you most recently adopted on your farm? CODE: 1-seed/Variety, 2-Agro-chemicals, 3-New animal breed, 4agronomic practices, 5-soil or water conservation, 6-conservation agriculture, 7Machinery/tools, 8-Storage method, 9-mono0cropping, 10-drying/processing, 99Others (Specify) 3 What is the depth of ground water level in this area?(ft) 98 Don’t know 4 Over the past 10 years have you experienced fall in ground water level?(0- No, 98-Don’t Know, If YES, by how much ft water level has declined) If NO, skip to 6 5 If YES, In your opinion, what are the reasons for the decline in ground water level? CODE A 6 Are you aware of any water conservation practices? (1- Yes, 0-No) if NO, skip to next section 101 CODE 7 8 If yes, name the practices? (Record up to three) A (CODE B) Are you using any of these practices on your farm? CODE: 1-Yes 0-No 9 If NO, why not? CODE C B Code A: 1- Indiscriminate use of water, 2. Deforestation, 3- Increase in population, 4-Increased industrial activity, 5- Water Pollution, 6-Decline in rainfall, 7-Reason not related to human activity,8- Increase in submersible pump, 98-Don’t Know, 99Others (specify) Code B: 1-Scheduling irrigation only when required, 2. Planting less water requiring crops/variety, 3. Keeping residue for water conservation, 4- Adopting water saving technology, 5- Farm pond, 99. Others (specify) Code C: 1- Beyond my control, 2- Single effort will not help, 3- Water saving technology are costly, 4- My land uses surface water, 5- No water problem in my area, 99- Others (specify) 5.2 Use of technology by farmers in social network 1. How many farmers in this village/other villages you know personally? 1-more than 100 farmers 2. 75-100 farmers 3. 50-75 farmers 4. 30-50 farmers 5. 20-30 farmers 6. less than 20 farmers, 100-No Response a=Zero Tillage b=LLL 2. Approximately how many farmers YOU KNOW c=DSR have used/currently using the following d=Residue retention/mulching technologies on their farm? (read each, and note e=legume rotation the numbers) f=drip irrigation Code: g=green seeker 997--I am not aware of this practice myself h=leaf color chart 999--don’t know how many are using it i=Nutrient expert decision support software j=agroforestry 5.3 Constraints to Wheat production S.No Questions 1a What are the two main constraints you face in wheat First 1b farming? CODE A Second 2 Have you stopped planting any wheat varieties in the past 3 years that you used to grow before?(1-Yes,0-No) If NO, skip to 5.4 3 If yes, How many? 4 Name the most recent variety you have discontinued 5 What is the main reason for discontinuation? CODE B Code A: 1-Land, 2-labour, 3-cash constraint, 4-seeds not available, 5-insect and disease problem, 6-cannot sell the crop, 7-price too low, 8-no information or technical advice on farming practices, 9-introduction of new variety, 99-other(specify), 100-No Response Code B: 1-Seed not available, 2-had low yield, 3-did not like the color, 4-susceptible to disease, 5- not liked by processors, 6-unpleasing cooking quality/taste, 99-other (specify) 5.4 Constraints to Rice production S.No Questions 1a What are the two main constraints you face in rice First farming? CODE A 1b Second 2 Have you stopped planting any rice varieties in the past 3 years that you used to grow before?(1-Yes,0-No) If NO, skip to 5.5 3 If yes, How many? 102 C 4 5 Name the most recent variety you have discontinued What is the main reason for discontinuation? CODE B Code A: 1-Land, 2-labour, 3-cash constraint, 4-seeds not available, 5-insect and disease problem, 6-cannot sell the crop, 7-price too low, 8-no information or technical advice on farming practices, 9-introduction of new variety, 99-other(specify), 100-No Response Code B: 1-Seed not available, 2-had low yield, 3-did not like the color, 4-susceptible to disease, 5- not liked by processors, 6-unpleasing cooking quality/taste, 99-other (specify) 5.5 Loss Due to Unexpected Weather 1. In the past 2 years, have you ever lost a significant portion of your wheat or rice production due to unexpected weather (eg low rainfall, flooding, unexpected monsoon time, hail, etc)? 1-Yes, 0No 2. If yes, how many times you have suffered such losses in past two years? 98-Don’t know/can’t say/don’t remember 5.6Access to information, infrastructure and credit 1 2 3 4 5 6a 6b 7 8 9 10 11 12 13 14 15a What is the distance from your house to the nearest paved road (if the house is next to the paved road, write zero) km What is the distance from your house to the nearest market where you obtain agricultural inputs (e.g., fertilizer, pesticides, seeds, etc.) km What is the distance from your house to the nearest agricultural extension office km What means of transport do you mainly use to get to the nearest commercial town? 1- Walking, 2- Tractor, 3- Bicycle, 4-Motorcycle, 5-Car, 6- Bus, 7- Light transport Vehicle, 8Animal Driven Cart, 99- others (specify) Distance from your home to this commercial town km a. Hou rs Time it takes on average to travel to this commercial town using the main mode of transportation (Consider time of one way travel) b. min utes Do you have a bank account? Code: 1-Yes 0-No Do you own kisan credit card? (1-Yes, 0-No) Do you currently have crop insurance policy (other than KCC)?Code: 1-Yes 0-No, If YES, skip to 12 Did you have crop insurance in the past but have discontinued? Code: 1-Yes 0-No Reason for not have crop insurance:1=Do not need, 2=Not available in my place, 3= No claim available at time of damage 4=Too expensive, 99= Others (Specify) Did you access credit for agricultural production in the past 12 months (1=Yes, 0= No) If NO, skip to 14 If yes, from where? 1- Bank, 2- Cooperatives, 3- SHG, 4-Community member, 5-Relative/ Friend/ Neighbour, 6Local money Lender, 7-Commission Agent, 8- Employer, 9-Agrovet, 10-Trader 99- Others (Specify) If no, why not? 1= Did not need, 2= Do not have access to credit, 3=Very high interest rate , 4-Far From Residence, 5- Bank staff not cooperative, 6- was getting less amount 7-loan was not approved 8-no collateral 99= Others (Specify) During the past year, where did you receive most of your information and advice about wheat production and marketing from? CODE A (Multiple Response) 103 15b If yes in previous question, In total, approximately how many times did you receive information about wheat last year from all these sources? 16a During the past year, where did you receive most of your information and advice about Rice production and marketing from? CODE A (Multiple Response) 16b If yes in previous question, In total, approximately how many times did you receive information about Rice last year from all these sources? 17 Do you use mobile phone to access information related to farming? 1=Yes 0=no If NO, End the Survey 18 If yes, what type of information? CODE C – Multiple Response 19 From whom do you access information using mobile phone? CODE D— Multiple Response CODE A. 1-Extension agent, 2-NGO staff, 3-Trader / input dealer, 4-Farmer group/leader farmer, 5-Service Provider, 6-I did not receive any information or advice, 99-other (specify) CODE B. 1=1-2 times, 2=2-3 times, 3=3-5 times, 4=5-10 times, 5=More than 10 times CODE C. 1-Weather, 2-Price, 3-Inputs, 4-Production technology, 5-Pest control, 6-government programs, 99other (Specify) CODE D. 1-Relatives/friends, 2-Input dealers, 3-KVK, 4-Kisan Call Center, 5-Extension agents, 6-RML, 7-IFFCO Kisan Sanchar Limited (IKSL), 8-mKRISHI, 9-other mobile based agro advisory services, 99-Other (specify) To be filled by Enumerator after the completion of Survey End Time Was the survey completed in first attempt or required a revisit? 1-First Attempt, 2-Re-Visit 104 APPENDIX D: LEA Implementer 3 Questionnaire Q.No . Scree n No. Question 1 1 State 2 1 District 3 1 Block 4 1 Village Options SECTION-A: LOCATION Bihar Haryana Vaishali Karnal Questio n Type Looping Drop Down None Drop Down Drop Down Alpha Numeric None None None SECTION-B: FARMER PROFILE 5 2 Name Alpha Numeric None 6 2 Age Numeric None Drop Down None 7 2 Gender 8 2 Education 2 Which of these do you have? 9 10 2 Marital Status 11 2 Education (Spouse) Male Female Illiterate Literate but no formal schooling/ School Upto 4 years School- 5 to 9 years SSC/ HSC Some College (a Diploma) but not Grad Graduate/ Post Graduate: General Graduate/ Post Graduate: Professional UID BPL Card Co-ooperative Society Membership Married Unmarried Illiterate Literate but no formal schooling/ School Upto 4 years 105 Drop Down None Multiple Check box None Drop Down Drop Down If "Married", go to 11, if Ünmarried", go to 13 None 12 13 2 3 14 15 3 3 16 3 17 3 18 4 19 4 20 4 Occupation (Spouse) School- 5 to 9 years SSC/ HSC Some College (a Diploma) but not Grad Graduate/ Post Graduate: General Graduate/ Post Graduate: Professional Farming Salaried Employment Casual Labourer-on farm Casual Labourer-off farm Self-employed off farm Housewife SECTION-C: HOUSEHOLD DETAILS Yes Are you head of household No Name of Head of Household Age of Head Illiterate Literate but no formal schooling/ School Upto 4 years School- 5 to 9 years SSC/ HSC Education of Head Some College (a Diploma) but not Grad Graduate/ Post Graduate: General Graduate/ Post Graduate: Professional Spouse Son/Daughter Relationship with the Parent head Brother/Sister Son-in-law/Daughter-in-law Grandchild Total family members in household Family Members Aged 0-17 Male Members working in agri 106 Drop Down None Drop Down If "Yes", go to 18, if "No", go to 14 Alpha Numeric Numeric None None Drop Down None Drop Down None Numeric None Numeric None Numeric None 21 4 22 4 23 4 Male Members working in non-agri Female Members working in agri Female Members working in non-agri Numeric None Numeric None Numeric None If "Yes", go to 25, if "No", go to 27 Yes 24 4 Migration in past 12 months Drop Down No 25 4 26 4 Reason for migration 27 5 Type of House 28 5 Source of energy for cooking 5 Household Items owned 29 30 5 31 5 32 5 33 5 34 5 35 5 Alpha Numeric Place of Migration To earn higher wages from farm activity To earn higher wages from non-farm activity To reduce burden on family Non-availablity of work in village To work on relative's farm Other Pucca Kuchcha Semi-Pucca Firewood and chips, charcoal or none LPG Other TV Two wheeler Pressure cooker Electric Fan Almirah/Dressing Table Sewing Machine Total Annual Household Income (Rs.) Total Annual Income from Agriculture (Rs.) Percentage of Income from Wheat Percentage of Income from Rice Wheat growing experience (Years) Rice growing Experience (Years) SECTION-D: LAND PROFILE 107 None Multiple Check box None Drop Down None Drop Down None Multiple Check box None Numeric None Numeric None Numeric None Numeric None Numeric None Numeric None 36 6 37 6 38 6 39 6 Distance to paved road Distance to agri-input market Distance to agriextension centre Last Time you adopted a new technology (YYYY) 40 6 Type of Technology Adopted 41 6 Total area owned (in local units) 42 6 All crops cultivated in past 12 months 43 6 Technologies used in Rabi 2014-15 6 Technologies used in Kharif 2015 44 Seed/Variety Agrochemical New animal breed Agronomic Practice (Weeding/Planting/Harvestin g) Soil/water conservation Machinery/Tools Storage Method Monocropping Drying/Processing Wheat Rice Cotton Sugarcane Maize Vegetables Other Laser Land Leveller Zero Tillage Direct Seeded Rice Laser Land Leveller Zero Tillage Direct Seeded Rice Numeric None Numeric None Numeric None Numeric None Multiple Check box None Numeric None Multiple Check box None Multiple Check box None Multiple Check box None Drop Down Üsed it and still using it"will go to 49, üsed it but discontinue d will go to 48 and skip to end of SECTION-E: TECHNOLOGY X13 Used it and still using it Used it but discontinued 45 7 Ever used TECHNOLOGY X Never used it 13 Repeated for each of the three technologies in the original questionnaire 108 Section , never used it will go to 46, 47 and skip to end of Section 46 47 48 7 Reasons for not using 7 Will you use if there is access (thru service provider) 7 49 8 50 8 51 8 52 8 53 8 Reasons for discontinuing Unwilling to take risk Expensive to hire Non-availablity in village Less Yield Not Satisfied High Weed Not suitable for small farmers Not suitable for crop Does not have irrigation facility Land is naturally level Cannot Say Yes No Unwilling to take risk Expensive to hire Non-availablity in village Less Yield Not Satisfied High Weed Not suitable for small farmers Not suitable for crop Does not have irrigation facility Land is naturally level Cannot Say When did you use for first time When did you use last time Who was involved in decisions? Only self Self and spouse Only male members of the family Whole Family Area cultivated using TECHNOLOGY X Crops cultivated using TECHNOLOGY X Wheat Rice Maize 109 Multiple Check box None Drop Down None Multiple Check box None Numeric None Numeric None Multiple Check box None Numeric None Multiple Check box None 54 8 55 8 56 8 57 8 58 8 59 8 Cotton Sugarcane Vegetables Other Local Government officials Private Input dealers Local NGOs / Development Agency Print Media (Newspaper / Source of information Magazines) on TECHNOLOGY X Electronic Media (TV /Radio) Local Exhibition / Agricultural Fair Fellow farmers / Relatives Local research Institute / KVK / ARS Save Water Soil management Improves crop yield More uniform moistureenvironment for crops Less time and water required Benefits derived from in irrigation TECHNOLOGY X Reduce weeds problem Easy land preparation Improves uniformity of crop growth and maturity Reduce consumption of seeds, fertilizers, chemicals, fuel, labor Increase Effect on male Labour Decrease Neutral Increase Effect on female Decrease Labour Neutral Yes Training Received No Multiple Check box None Multiple Check box None Drop Down None Drop Down None Drop Down If Yes, go to 59, if No, go to 60 Alpha Numeric Source of Training Owned 60 61 9 Owned/Leased 9 Cost of Hiring (Rs/Land Unit or Rs/hr) Leased Alpha Numeric Numeric 110 None If Leased, go to 61, if owned, go to 64 None 62 9 Source of Hiring 63 9 Age of Operator 64 9 Gender of Operator 65 66 9 9 67 9 Purchase Price (Rs.) Year of Purchase Amount of subsidy received (Rs.) Service provider in village Service provider in other village Village co-operative Relatives/Neighbours Farmers Association Progressive farmer Drop Down Numeric Drop Down None Numeric Numeric None None Numeric Numeric None If "Yes", go to 69, if "No", go to 71 None Numeric None Numeric None Numeric Lack of awareness Lack of availability of laser leveler High cost of use Multiple Reason others not Farmers’ are not perceiving Check using benefit out of laser leveler box Lack of availability of Driils High cost of use Farmers’ are not perceiving benefit out of DSR SECTION-H: AGRICULTURAL INFORMATION AND CREDIT Fellow Farmers Local Input Dealers Govt. Extension officers KVK/ARS/SAUs Multiple Source of infromation Local NGO Check on seed variety/agri box Radio/TV input Newspapers/Magazines Exhibitions/Agri Fair Mobile Agro Advisory Services RML, IKSL, mKrishi None Male Female Yes 68 10 69 10 70 10 71 10 72 10 73 10 129 19 None Do you lease out TECHNOLOGY X? Drop Down No Last time leased out Revenue earned (per unit) Fellow Farmers known Farmers using TECHNOLOGY X on your advice 111 None None None 130 131 132 19 Source of infromation on agricultural technology 19 Source of infromation on agricultural markets and commodity prices 19 Use mobile phone for agri information 133 19 Type of information accessed 134 19 Source of information access 135 20 Do you have Kisan Credit Card (KCC) 136 137 20 20 Fellow Farmers Local Input Dealers Govt. Extension officers KVK/ARS/SAUs Local NGO Radio/TV Newspapers/Magazines Exhibitions/Agri Fair Mobile Agro Advisory Services RML, IKSL, mKrishi Fellow Farmers Local Input Dealers Govt. Extension officers KVK/ARS/SAUs Local NGO Radio/TV Newspapers/Magazines Exhibitions/Agri Fair Mobile Agro Advisory Services RML, IKSL, mKrishi Yes No Weather Price Input use Production Technology Pest control Govt. Programs Relatives/ Friends Input Dealers KVK Kissan Call Center Extension Agents RML IFFCO Kisan Sanchar Limited (IKSL) mKRISHI Other mobile based agro advisory services Yes No Credit Limit on KCC (Rs.) Yes 112 Multiple Check box None Multiple Check box None Drop Down If Yes, go to 133, if No, go to 135 Multiple Check box None Multiple Check box None Drop Down If Yes, go to 136, if No, go to 137 Numeric None 138 20 Have you taken any loan in past 12 months Amount of loan (Rs.) 139 20 Sources of loan 140 20 Purpose of loan 21 No. of crop losses due to weather since 2013 141 Drop Down No Bank Commission Agents / Aadati Relatives / Friends To buy new tractor / machinery To procure agri-inputs (Fert. / chemicals) Irrigation systems Investment in field (Land levelling etc.) Function at Home Purchase of vehicle Numeric Multiple Check box Multiple Check box Numeric Yes 142 21 143 21 144 22 145 22 146 22 147 22 Use crop Insurance No Not available in my area Reason for not using Too Expensive crop insurance Other SECTION-I: CONSTRAINTS Land quality Labor constraint Cash constraint Seeds not available Two important Insect and Disease problem constraints faced in Cannot sell the crop wheat farming Price too low No information or technical advice on farming practices None/no more Number of wheat varieties discontinued in past 3 years Name of one wheat variety discontinued Seeds not available Had low yield Reason for Did not like the discontinuing the shape/size/color variety Susceptible to diseases Not liked by processors 113 Drop Down If Yes, go to 138, if No, go to 141 None None None None If No, go to 143, if Yes, go to 144 Drop Down None Multiple Check box None Numeric Alpha Numeric Multiple Check box None None None 148 22 149 22 150 22 151 22 152 23 153 23 154 23 155 23 Two important constraints faced in rice farming Number of rice varieties discontinued in past 3 years Name of one rice variety discontinued Plot Landmark Plot Area (Local Units) 156 23 157 23 Soil Type 14 23 Irrigation Status Multiple Check box Numeric Alpha Numeric Seeds not available Had low yield Reason for Did not like the discontinuing the shape/size/color variety Susceptible to diseases Not liked by processors SECTION-J: PLOT INFORMATION14 1 2 How many plots do 3 you have 4 5 Total Area cultivated across all plots (Local Units) Plot-1 Ownership Status 158 Land quality Labor constraint Cash constraint Seeds not available Insect and Disease problem Cannot sell the crop Price too low No information or technical advice on farming practices None/no more Owned Leased Shared Sandy Sandy Loam Clayey Loam Clay Rainfed Irrigated This is repeated for up to five plots in the original questionnaire 114 None None None Multiple Check box None Multiple Check box None Numeric Alpha Numeric None Numeric None Drop Down None Drop Down None Drop Down If Irrigated, go to 159, if rainfed, go to 16 None 159 23 Source of Irrigation 160 23 Method of Irrigation 161 24 162 24 163 24 164 165 24 24 Wheat Intercropping Percentage planted to wheat Total Wheat Production (Quintals) Total Quantity sold (Quintals) Selling Price (Rs./Qtl) 166 25 Method of Plot Preparation 167 25 Method of Levelling 168 25 Method of seeding 169 25 170 25 Canal Pond Well/Tubewell Flood Drip Sprinkler Pivot Yes No Planker Tiller/ Cultivator Rotavator Harrow ZT drill Laser leveller Conventional ploughing None Traditional Land Levelling Laser Land Levelling Broadcasting Seed cum ferti drill ZT drill Turbo happy seeder Seed Rate (kg/Land Unit) Wheat Variety Planted 171 25 Seed source 172 25 Labour Provided by 173 25 Gender of Labour Drop Down None Drop Down None Drop Down If Yes, go to 162, if No, go to 163 Numeric None Numeric None Numeric Numeric None None Drop Down None Drop Down None Drop Down None Numeric Alpha Numeric Own harvest Purchased Received free or subsidized from govt/NGO Borrowed from other farmer Don’t remember Self Spouse Both Son/Daughter Other Male 115 None None Multiple Check box None Multiple Check box None None Female Drop Down Drop Down If Yes, go to 175, if No, go to 176 Numeric None Numeric None Numeric Numeric None None Drop Down None Drop Down None Drop Down None Yes 174 26 175 26 176 26 177 178 26 26 Rice Intercropping Percentage planted to Rice Total Rice Production (Quintals) Total Quantity sold (Quintals) Selling PRice (Rs./Qtl) 179 27 Method of Plot Preparation 180 27 Method of Levelling 181 27 Method of seeding 182 27 Seed Rate (kg/Land Unit) 183 27 Rice Variety Planted 184 27 Seed source 185 27 Labour Provided by 186 27 Gender of Labour 187 28 Do you have another plot? No Planker Tiller/ Cultivator Rotavator Harrow ZT drill Laser leveller Conventional ploughing None Traditional Land Levelling Laser Land Levelling Broadcasting Transplanting Seed cum ferti drill ZT drill Turbo happy seeder Numeric Alpha Numeric Own harvest Purchased Received free or subsidized from govt/NGO Borrowed from other farmer Don’t remember Self Spouse Both Son/Daughter Other Male Female 116 None Multiple Check box None Multiple Check box None Drop Down None Drop Down Yes None If Yes, go to 188, If No, end of survey APPENDIX E: CEA Implementer – Wheat and Rice Questionnaire To be filled by enumerator Date of the interview (dd/mm/yyyy) Name of the enumerator Time started To be filled by Supervisor Date checked (dd/mm/yyyy) Name of the supervisor Section 1: Current household composition and characteristics 1.1 Household identification 1a District CODE A 1b. Mandal / Block 2a Village/hamlet name 3 Indicate random selection of Household Elevation (in meter) GPS coordinate N (Format xx.xxxxx) Household (HH) id 4 4 6 1c. Gram Panchayat 2b. CODE CODE B District Code GPS coordinate E (Format xx.xxxxx) Mandal / Block Village Code Code HH Number Code A: 1-Karnal, 2-Ludhiana, 3-Vaishali, 4-Kurnool, 5-Anantapur Code B: 1-Original, 2-Replacement 1.2 General information about the Respondent Note: Respondent here refers to the Lead Decision Maker for Agricultural Activities in the family. 1 Name of the Respondent a. First Name b. Last Name 2 3 4a 4b 4c 5 6 7a 7b 8 9a 9b 9c 10 11 12 Gender of the Respondent (Code: 1 –Male, 2- Female) Age of the Respondent Marital status (CODE A) How many brothers and sisters do you have (including siblings that may a. b. sisters have died)? brothers What was your birth order? For example, were you first born, second born, Years third…? of formal education of Respondent and spouse (if a. Respondent b. Spouse (if married) married) Main occupation (CODE B) a. Respondent b. Spouse (if married) Can read a local Indian language (1-Yes, 2-No) Can read English (1-Yes, 2-No) Years of Experience in farming Years of experience in growing groundnut Years of experience in growing wheat Years of experience in growing rice Mobile number Relationship with Head of the household (HOH) (CODE B) If option 1, skip to question 17 Name of the Head of Household 117 13 14 15 Gender of the Head of the HH (Code: 1 –Male, 2- Female) Age of the Household Head Years of formal education 16a 16b 17a a. Head b. Spouse of Head (if married) Can read a local Indian language (1-Yes, 2-No) Can read English (1-Yes, 2-No) Religion of the household (Code: 1- Hindu, 2- Muslim, 3-Christian, 4- Sikh, 5-Buddhist 98-Don’t know, 99Others (specify) 17b Caste (Code 1- General, 2- SC, 3- ST, 98-Don’t Know, 99 –Others) 18 Highest level of formal education completed by any a. b. Gender (1-Male 2member of the household (Years of education), and the Education Female) gender of that individual 19 Are you a member of any farmer organization or a farmer cooperative (1-Yes 2-No) 20 If Yes, what is your level of involvement in this group? 1- very active, 2-somewhat active, 3-not active 21 Are you a leader of any of these groups? (1-Yes 2-No) 22 How many sons and daughters do you have? a. Sons b. Daughters 23 If farmer has both sons and daughters ask: What is the birth order of your nd eldest son? Is he 1stspouse, born, 2 born,…? Code A: 1-Married living with 2-Married but spouse away, 3-Divorced, 4-Widow, 5-Not married, 99-other (specify), Code B:1- Farming on own farm, 2- Livestock rearing, 3- Salaried employment, 4- Self-employed off farm, 5- Casual labourer on-farm, 6- Casual labour off farm, 99-other (specify). Code C: 1-Head himself/herself, 2- Wife, 3- Husband, 4- Son, 5-Daughter, 6- Grandchild, 7- Father, 8-Mother, 9- Sister, 10Brother, 11- Niece, 12- Nephew, 13- Son in law,14- Daughter in law, 15-Brother in law, 16-Sister in Law, 17- Father in law, 18- Mother in law, 19- Other family relatives, 20- Servant, 21- Permanent labour, 22-Tenants, 99- Other person not related 1.3. Household Information A 'household' is usually a group of (related or unrelated) persons who normally live together and take their meals from a common kitchen. If a group of unrelated persons live in a census house but do not take their meals from the common kitchen, then they are not constituent of a common household. 1.3.1 How many members belong to this household: _______; By age and gender FEMALE: a. <5 years ____ b. 5-17 ______ c. >18 _____ d. Total female members_____ MALE: e. <5 years ____ f. 5-17 ______ g. >18 _____ h. Total male members_____ 1.3.2 Total working members in the family (in Numbers) a. Male________ b. Female _________ 1.3.3 In the past 12 months, did any member of your household obtain income from any of the following sources? (Instruction: Read each item and note yes/no) 1=Yes 0=No g. Sale proceeds of Field Crops h. Horticulture crop sales i. Dairy j. Livestock sales for meat k. Renting/leasing land or farm equipment l. Wages from farm labor 1=Ye s g. Wages from off farm (govt. job, teacher,etc) h. Non-farm business or selfemployment i. Remittance j. Pension Income k. Other (specify) 1.3.4 Total annual household income across all the activities and working members (Rs) a. Cash: (CODE A) b. in kind (cash equivalent) (CODE A) 118 0=No Code A: 0 = < 25,000 1= 25,000-50,000, 2=50,000-1,00,000, 3= 1,00,000- 2,00,000, 4= 2,00,000-3,00,000, 5= 3,00,000-4,00,000, 6= 4,00,000-5,00,000, 7= 5,00,000 – 6,00,000, 8= 6,00,000 – 8,00,000 , 9= 8,00,000 – 10,00,000, 10= greater than 10,00,000 1.3.5. What source of income mentioned above contributes the largest share to your total household income? (write a code a to k corresponding to the source mentioned) ________ 1.3.6 In your estimate, what percentage of your total HH income in the last 2 years came from? a. wheat farming ________ b. rice farming _________ 1.3.6 What type of Ration card do you have? __________ 1-APL (white), 2-BPL (blue), 3-AYY (yellow), 4-AY (special), 5-Do not have ration card, 99-Others (Specify) 1.3.7 Is anyone in your household (other than you) a member of a farmer producer organization or a farm cooperative?__________ 1-Yes, 2-No 1.4 Poverty Score Card (The codes correspond to the poverty SCORE, PLEASE KEEP THESE SCORE CODES when programming the survey) 1 2 3 4 5 6 7 8 9 10 How many people aged 0-17 are currently part of your household? 0-Five or more, 4-Four, 8-Three, 13-Two, 20-One, 27-None What is household’s principal occupation? 0-Laborers (agricultural plantation, others farm), hunters, tobacco preparers and tobacco product makers and other labourers 14-Professionals, technicians, clerks, administrators, managers, executives, directors, supervisors and teachers 8-Others Is the residence all pucca (burnt bricks, Stone, Cement, Concrete, Jack board/Cementplastered reeds, timber, tiles, gal vanished tin or asbestos cement sheets)? 4-Yes, 0-No What is the household’s primary source of energy for cooking? 0-Firewood and chips, Charcoal or none, 17-LPG, 5-Others Does the household have own television? 6-Yes, 0-No Does the household own a bicycle, scooter or motor cycle? 5-Yes, 0-No Does the household own an almirah/dressing table? 3-Yes, 0-No Does the household own a sewing (tailoring) machine? 6-Yes, 0-No How many pressure cookers or pressure pans does the household own 0-None, 6-One, 9-Two or more How many electric fans does the household own? 0-None, 5-One, 9-Two or more 1.5. Information on Migrant Family Member (Enumerator: A member is usually termed as migrated if she/he lives outside village for more than a year or left recently with that intention) Only for related family members .Exclude spouses /children of migrant members) 1a 1b Has any member of your household migrated in the past 5 years? (1Has any member of your household migrated in the past 12 months? Yes, 2-No) If NO to both these questions, skip to next section (2.1) (1-Yes, 2-No) 2 Place of most recent migration CODE A 3 Reasons for migration CODE B 4 Does the member who has migrated take major decisions in matter relating to the agricultural activities? (1-Yes, 2-No) 5 Does the member who has migrated contribute towards meeting household Code A: 1-Within state (urban 2- Within state (rural area), 3-Within country, another state, 4-Middle East, 5expenses? (1-Yes,area), 2-No) US/Canada/Australia, 6-European Countries, 99. Others (specify) Code B: 1- Better prospect of employment, 2- Weather related uncertainties, 3- Higher education, 4- Marriage, 98- Don’t know, 99-Others (sp) Section II. Land Holding 2.1a Local land unit (LU) CODE A 119 2.1b 2.1c 2.1d 2.1e 2.1g 2.1f LU conversion rate How many Plots of land does your household own? How much area of land your household owns across all these plots?much (LU) land that your household currently owns was: How Acquired through purchase? Inherited? Acquired through other means (specify)? 1 acre =...................LU Total across 2.1e to 2.1g should equal 2.1d 2.1h Have you ever sold any land that you had inherited? If No, write 0, If Yes, indicate the total land area sold ; 999-did not inherityou anyever landsold any land that you had acquired through purchase? If No, 2.1i Have write 0, If Yes, indicate the total land area sold; 999-have not purchased any 2.1j If YES to either 2.1h or 2.1i: What was the main reason for selling the land? land CODE B 2- Acre, 3- Killa, 4- Kanal, 5-Bissa, 99- Others (specify) Code A: 1- Bigha, Code B: 1-to pay off debt; 2- to get cash for non-farm business or investment; 3-to meet household expenses; 4-to downsize my farming operation; 5-Other (specify) III. Technology specific questions 3.1. Laser Land Leveller S. Questions N 1 2 3 4 5 6 7 8a 8b 9 10 11 12 13 14 15 a 15 Have you ever heard about LLL or laser land leveller? (1-Yes, 2-N0) If NO, skip to When did you first come to know about it? (9999-Don’t Know) YYYY next section (3.2) Source of information CODE A Have you ever used LLL? (1-Yes, 0- No) If YES, skip to 7; if NO, ask 5 and 6 and skip next section If No,towhy? CODE B(3.2) (main reason) Will you adopt it if there is access to a service provider? (1-Yes, 0- No, 98-Don’t know/can’t say) When did you start using it? (9999-Don’t Know) YYYY Who was involved in making the decision to use LLL? CODE C What was the motivation behind the decision to use LLL? CODE D Did you stop using it once you adopted it?(1-yes, 0-No) if NO, skip to 11 If yes, why? (main reason) CODE B Have you or anyone in your household received training in using LLL? (1-Yes, 0No) If NO, to 13 (CODE A, 1 to 8) If Yes, Fromskip whom? When was the last time you used LLL (indicate c. Season (1-Kharif 2d. Year season Did youand useyear) your own leveler or hire it? (code 1-own 2-Hired) If own, skip to Rabi) 16; If hired, ask 15 and then skip to 23 What was the per unit cost of hiring LLL Select the Unit when you leveled last? 1=hour 2=acre 99=other (specify) Rupees per unit b 15 From whom did you hire the LLL? CODE E c 16 At what price did you purchase the LLL Machine? (Rs) 17 When did you purchase it? (YYYY) 120 18 19 20 21 a Did you receive any subsidy at the time of purchase of ANY?(0-No, if YES amount of subsidy) Do you lease this machine to others on rental basis? 1-Yes 0-No, If NO, skip to 23 When was the last time you rented the LLL to others? YYYY What was the per unit revenue you earn from renting out the LLL when you rented to others last time? 21 Select the Unit 1=hour 2=acre 99=other (specify) Rupees per unit you charged b 22 For how many units did you rent out your LLL last time you rented to others? 23 Who mainly operates LLL on your farm? Capture age and gender a Age 23 Gender 1=Male 2=Female b 24 25 26 27 28 29 30 31 31 a 31 b 31 c 31 d e On how many plots on your farm did you use this technology in the last season you used What is theit? total area cultivated using this technology in the last season you used it? LU What crops were cultivated on the plot in which you used this technology a B in the last season you used it? CODE F In your opinion, what are the main benefits of using laser land leveler? 1-Uniformity of crop growth / maturity 2-Reduce water requirement / saves irrigation water and cost 3-Improves crop establishment 4-Increases water application efficiency 5-Higher Yields Do you face any inconveniences in using LLL? (Rank Top 3) 1-Unavailable at the peak time 2-too expensive to hire 3-Service provider does not provide credit 4-Lack of service provider in the village 5-Unsatisfied with technology 6-Unsatisfied with service quality Rank Top 3 6-Increase nutrient efficiency 7-Reduces weed problem 8-Labor saving 99-Others (Specify) 97-No more benefits 7-Difficulty in getting subsidy 8-lack of repair and service fertility nearby 9-Frequent technical problems with Machine 99-Others (Specify) 97-No inconvenience Do you share your LLL experience with other farmers? (1- Yes, 0-No) In your opinion, does the laser land leveller increase, decrease or has no effect on the time devoted to farming by the MALE members of your household relative to the conventional practice? (Code 1Increase, 2-Decrease, 3-Neutral)If Neutral, skip to 32 If response is increase or decrease: In what aspects land is the labor input by MALE members increased or preparation Sowing decreased? Weeding 1-Yes, 0-No Irrigating Harvesting 121 Rank Top 3 c d 31 32 f Other (specify) In your opinion, does the laser land leveller increase, decrease or has no effect on the time devoted to farming by the FEMALE members of your household relative to the conventional practice? (Code: 1-Increase, 2-Decrease,3-Neutral)If Neutral, skip to Section 3.2 If response is increase or decrease: In what aspects land is the labor input by FEMALE members increased or Sowing preparation decreased? Weeding 1-Yes, 0-No Irrigating Harvesting Other (specify) 33 33 a 33 b 33 c 33 d 33 e Code f A: 1-Government Extension service, 2- Service Provider, 3- CIMMYT/ICRISAT, 4- Farmer Cooperative /group, 5-Research centres other than CIMMYT/ICRISAT, 6- Neighbour/Relative farmer, 7- Private Company/input dealer, 8- NGO/CBO, 9- Radio, 10- TV, 11- Mobile Phone , 12- Newspaper,, 13-traditionally known, 99- Others (Specify) Code B: 1-Unwilling to try new technology, 2-lack training/information, 3- Expensive to hire/build, 4- Service/materials not available in the village, 5-Gives Less Yield, 6- Not satisfied with output, 7- Does not look good, 8- High weed, 9- Not suitable on small Land, 10- Not suitable for the crop, 11- Does not have irrigation facility, 12- Land is naturally level/ no need, 13-Difficulty in getting subsidy, 14- Lack of information 98-Cannot say, 99- Others (Specify) Code C: 1-Only I myself made the decision, 2-Both me and my spouse were involved, 3-I and other male members of my family made the decision, 4-Whole family was involved, 99-Others (Specify) Code D: 1-to increase crop yield/productivity, 2-to reduce irrigation cost or water wastage, 3-to control weed problem, 4other farmers in the village were using it, 5-Other (specify) Code E: 1-Service provider in village, 2- Service provider from other village, 3-Village cooperative, 4-Relatives/Neighbour farmer, 5-Farmers association, 6-Progressive farmer, 7-Farmer Cooperative, 99-Others (Specify) Code F: 1-Rice, 2-Wheat, 3-Pulses, 4-Vegetables, 5-Fodder, 99-Others (Specify) a IV. Plot characteristics and wheat/rice production in Rabi 2014-15 and Kharif 2015 season 4.1 Land use 1a How many plots did you cultivate in Rabi 2014-15 season? 1b What was the total cultivated land area in Rabi 2014-15 season (LU) 2a How many plots did you cultivate in Kharif 2015 season? 2b What was the total cultivated land area in kharif 2015 season (LU) 3 Did you leave any land fallow in Rabi 2014 and this Kharif 2015 season? 1-Yes 2-No (If NO, go to 4.2) 4 Reason for leaving land fallow (1- Land not fertile, 2- Unavailability of water, 3Dispute over land, 4-Unabvailability of labour, 99-Others (Specify) What crops did your HH produce in the last 12 months in the following categories (only record number of crops mentioned): 5a. Cereal 5b. Pulses 5c. oil 5d. horticulture 5e. fibre crops 5f. Other crops 4.2. Plot characteristics For each of the plot cultivated in Rabi 2014-15 and Kharif 2015 season, I would like to ask you some specific questions. 122 S.N Note to Enumerators: Start with biggest plot farmer cultivated Plot I Plot 2 1a 1b 2 3a 3b 3c 4 5 6 7 8 9 10 11 12 13a 13b 14a 14b 14c 14d 15a 15b 15c 15e 15f Distance from your house to this plot km Size /Area of plot (LU) Plot ownership CODE: 1- Owned, 2.-Leased in /Shared in, 3Leased out/Shared OutCODE A If owned, Owned by If owned, was this plot inherited or purchased? 1-inherited 2purchased 98-don’t know Your assessment of the market value of this plot if you were to sell it today? Rs Irrigation source CODE: 0- No irrigation, 1-Tube Well, 2-Open Well, 3- River canal water, 4-Pond 99- Others (specify) Irrigation type CODE: 1-Flood (with pump), 2- Flood (without Pump), 3-Furrow, 4-Drip, 5-Sprinkler, 99-Other (specify) Soil type CODE: 1-Sandy, 2-Sandy Loam, 3-Loam Soils, 4-Clay Loam, 5- Clay, 99-Others (Specify) Soil quality CODE:1- Good, 2-Medium, 3-Poor Soil Salinity CODE:1- High, 2- Medium, 3- Low 98-don’t know Land level CODE:1- High level, 2- Middle level, 3- Low level, 4uneven/mixed, 5-uniform 98-don’t know What is your observation about the soil quality, fertility on this plot compared to last 10 years? (Code 1. Declining, 2. Remain same, 3. Improving) How many times this plot has been leveled? If ZERO, ask 12 then go to 14a ; other than ZERO, skip to 13a Why this plot has never been leveled? CODE B (Multiple responses possible) When was the plot leveled last? (mm/yyyy) Method of leveling used CODE:(0 –Traditional, 1-Laser land levelling) What did you do with the crop residues on this plot at the end of Kharif 2014? CODES: 0-No residue was produced, 1-Retained in the field, 2Mulched, 3-Burnt it, 4-Used it as fodder or cooking fuel on-farm, 5-Sold it, 99-Other (specify) If retained or mulched: Percentage retained/mulched? What did you do with the crop residues on this plot at the end of Rabi 2015? CODES: 0-No residue was produced, 1-Retained/incorporated in the field, 2-Mulched, 3-Burnt it, 4-Used it as fodder or cooking fuel on-farm, 5-Sold it, 99-Other (specify) (multiple responses are possible) If retained or mulched: Percentage retained/mulched/incorporated? Soil bunds, Are you currently using the Field/boundary bunds following technologies on this Broad bed and furrow plot or more generally on your Contour bunds farm with direct impact on this Polythene mulching 123 Plot 3 S.N Note to Enumerators: Start with biggest plot farmer cultivated Plot I Plot Plot 3 2 15g plot? (Check all that apply) 1Nala plugs/RFDs 15h Yes, 2-No Sunken pits 15i Farm pond 15j Masonry check dams 15k Well recharge pits 15l Penning Sheep/Goat/Cattle 15m Others (Specify)…………………… 16a Do you have following types of Fruit trees 16b planted trees on this plot? Trees for firewood/fuel 16c Trees for soil fertility 16d CODE: 1-Yes 2-No Trees for commercial Code A: 1-Head himself/herself, 2- Wife, 3- Husband, 4-purpose Son, 5-Daughter, 6- Grandchild, 7- Father, 8-Mother, 9- Sister, 10Brother, 11- Niece, 12- Nephew, 13- Son in law, 14- Daughter in law, 15-Brother in law, 16-Sister in Law, 17- Father in law, 18- Mother in law Code B: 0-Don’t know what is Laser levelling / this service is not available here; 1- Financial constraint, 2- Used on trial basis, 3-Land is naturally level, 4- Not required for particular crop, 5-Small land size, 6- Land not empty/vacant, 7-leased land , 99-Others(Specify) Following questions are specific to wheat farming in Rabi 2014 (continue for the same plots) S.N Note to Enumerators: Start with biggest plot farmer cultivated Plot I Plot Plot 3 2 17 Was this plot cultivated with wheat in Rabi 2014-15? CODE: 1Yes 2-No If No, Skip to next plot or next season Was wheat inter-cropped? CODE: 1-Yes 2-No If NO, skip to 21 18 If YES, what was the total value of other crops harvested on this 19 plot in Rabi 2014? 98- don’t remember 20 What percentage of this plot was planted to wheat in the Rabi Rs season? 21 Name of the inter/mixed Crop CODE C 98=Not cultivated any inter/mixed crop crop rotation on this plot by planting a legume 22 Did you practice crop before or after wheat crop? CODE: 1-Yes 0-No 23 Method of Plot Preparation in Rabi 2014 CODE: 1-Planker 2- Tiller/Cultivator, 3- Rotavator, 4- Harrow, 5Paddy Harrow, 6-ZT drill 7-LLL 8-conventional ploughing 9ripping 10-ridging 99- others (Specify) 24 Method of seeding in Rabiup2014 CODE:1-Broadcasting,2-seed Record multiple response to 4 for each plot cum ferti drill,3-ZT drill, 4-Turbo happy seeder 25 Seeding rate (kg/LU) 26 a. a. Relationship (CODE Who in your house mainly provided A) labour for this plot in Kahrif 2015? (write the relationship with the 27 b. b. Gender (1-Male, 2respondent and gender) Female) 28a Who is the main decision maker regarding inputs and outputs of this a. a. Relationship (CODE A) 124 28b plot? (write the relationship with the respondent and gender) 29 30 b. b. Gender (1-Male, 2Female) What wheat variety of seed was planted in Rabi 2014? Source of wheat seed planted CODE: 1=Saved from previous harvest, 2=Purchased from grain vendors in the market, 3=Purchased from other farmers or community based organization who produced seed, 4=Purchased from seed companies or input dealers, 5=Received subsidized seed from government or NGOs, 6=Received seed from extension agents, 7=Borrowed / obtained from neighbors/relatives, 98=Do not remember, 99=Others (specify) 31 Can you recall when was the first year you adopted this variety on your farm? 9999=Do not remember 32 What was the source of wheat seed for the first planting? CODE: 1=Purchased from market, 2=Purchased from other farmers or community based organization who produced seed, 3=Purchased from seed companies or input dealers, 4=Received subsidized seed from government or NGOs, 5=Received seed from extension agents, 6=Borrowed / obtained from neighbors/relatives, 98=Do not remember, 99=Others (specify) 33a What are the two characteristics of this variety First you LIKE? 33b Second CODE: 1=High yielding, 2= Resistance to Insect and disease, 3=Drought resistance, 4=Early maturity, 5=Seed quality, 6=Colour and taste, 7=Processing quality, 8=Good price / high demand, 99=Others (Specify) 34a What are the two characteristics of this variety First you DISLIKE?CODE: 1=Low yielding, 2=Susceptible 34b Second to insects and diseases, 3=Susceptible to drought, 4=Late maturity, 5=Seed quality, 6=Color and taste, 7=Processing quality, 8=Low price / Low demand, 99=Others (Specify); 97-No more 35 Did you use any of these a=organic fertilizer (kg) inputs / practices on this 36 b=Urea (Kg) plot in Rabi season(read 37 c=DAP (kg) each input and note the response) 38 d=Potash (Kg) 0-No, if Yes, note the total Quantity of input used) 39 e=Phosphate (Kg) 40 f= Zinc (kg) 41a g=pesticides 41b 42a h=herbicides 42b 43 i=hired labor (yes/no) 44 j=Nutrient Expert Decision Support software (yes/no) 45 k=Leaf Colour Chart (yes/no) 46 l=GreenSeeker sensors (yes/no) 47 m=other inputs (specify) 125 (choose unit mg/gm/ml/lit/kg) (choose unit mg/gm/ml/lit/kg) 48 49 50 51 52 53 54 55 56a 56b 57 What was the Total Man Days of Labor land preparation required in the following activities for wheat Sowing farming in Rabi Season 2014-15? Weeding Irrigating Harvesting Other (specify) How many times was this plot irrigated in Rabi Season 2014-15? What was the total quantity of wheat produced from this plot in Rabi 2014? Qtl sold Qtl Wheat quantity Price sold Rs/Qtl If plot is inter-mixed crop, ask: What is the total value of other crops you expect to harvest from this plot in this season? Code A: 1-Head Rs himself/herself, 2- Wife, 3- Husband, 4- Son, 5-Daughter, 6- Grandchild, 7- Father, 8-Mother, 9- Sister, 10Brother, 11- Niece, 12- Nephew, 13- Son in law, 14- Daughter in law, 15-Brother in law, 16-Sister in Law, 17- Father in law, 18- Mother in law Code B:1- Financial constraint, 2- Used on trial basis, 3-Land is naturally level, 4- Not required for particular crop, 5Small land size, 6- Land not empty/Vacant, 99-Others(Specify) Code C: 1=Red gram, 2=Green gram, 3=Horse gram, 4=Cow pea, 5= Black gram, 6= Bengal gram, 7= Lentils, 8=kidney beans / Rajma, 9=pigeon pea; 9= groundnut, 10=mustard, 11= Finger millet, 12= Little millet, 13= Foxtail millet, 14= Pearl millet, 15= Barnyard millet, 16=Kodo millet, 17= Proso millet, 18=Sorghum,19=Maize, 20-wheat, 21-rice, 22- soybean, 98=Not cultivated any inter/mixed/border crop, 99=Others (Specify) Following questions are specific to rice farming in Kharif 2015(continue with the same plots) S.N Note to Enumerators: Start with biggest plot farmer cultivated Plot I Plot 2 Plot 3 58 59 60 61 62 63 64 65 66a 66b Is this plot cultivated with rice in Kharif 2015? CODE: 1-Yes 2No If No, skip to Next plot or next section Is rice inter-cropped? CODE: 1-Yes 2-No If NO, skip to 61 What percentage of this plot is planted to rice in the Kharif season? Name of the inter/mixed Crop CODE C 98=Not cultivated any inter/mixed crop crop rotation on this plot by planting a legume Did you practice crop before or after the rice crop? CODE: 1-Yes 2-No Method of Plot Preparation in Kharif 2015 CODE: 1-Planker 2- Tiller/Cultivator, 3- Rotavator, 4- Harrow, 5Paddy Harrow, 6-ZT drill 7-LLL 8-conventional ploughing 9ripping 10-ridging 99- others (Specify) Record multiple response upto 4 forCODE:1-Broadcasting, each plot Method of seeding in Kharif 2015 2-seed cum ferti drill,3-ZT drill, 4-Turbo happy seeder Seeding rate (kg/LU) c. a. Relationship (CODE Who in your house mainly A) provided labour for this plot in Kahrif 2015? (write the relationship d. b. Gender (1-Male, 2with the respondent and gender) Female) 67a 67b c. a. Relationship (CODE Who is the main decision maker regarding inputs and outputs of this plot? (write the relationship with the respondent and gender) A) d. b. Gender (1-Male, 2Female) 126 S.N Note to Enumerators: Start with biggest plot farmer cultivated 68 What rice variety of seed is planted in Kharif 2015? Source of rice seed planted CODE: 1=Saved from previous harvest, 2=Purchased from market from grain vendors, 3=Purchased from other farmers or community based organization who produced seed, 4=Purchased from seed companies or input dealers, 5=Received subsidized seed from government or NGOs, 6=Received seed from extension agents, 7=Borrowed / obtained from neighbors/relatives, 98=Do not remember, 99=Others (specify) Can you recall when was the first year you adopted this variety on your farm?9998= less than 15 years ago but do not remember; 69 70 Plot I Plot 2 Plot 3 9999= more than 15 years ago but do not remember; 71 72a 72b 73a 73b 74 75 76 77 78 79 80a 80b 81a 81b 82 83 84 85 86 What was the source of rice seed for the first planting? CODE: 1=Purchased from market, 2=Purchased from other farmers or community based organization who produced seed, 3=Purchased from seed companies or input dealers, 4=Received subsidized seed from government or NGOs, 5=Received seed from extension agents, 6=Borrowed / obtained from neighbors/relatives, 98=Do not remember, 99=Others (specify) What are the two characteristics of this variety you First LIKE? Secon CODE: 1=High yielding, 2= Resistance to Insect and d disease, 3=Drought resistance, 4=Early maturity, 5=Seed quality, 6=Colour and taste, 7=Processing quality, 8=Good price / high demand, 99=Others (Specify) What are the two characteristics of this variety you First DISLIKE?CODE: 1=Low yielding, 2=Susceptible to Secon insects and diseases, 3=Susceptible to drought, d 4=Late maturity, 5=Seed quality, 6=Color and taste, 7=Processing quality, 8=Low price / Low demand, 99=Others (Specify) Did you use any of a=organic fertilizer (kg) these inputs / b=Urea (Kg) practices on this c=DAP (kg) plot in Kharif 2015 d=Potash (Kg) (read each input and note e=Phosphate (Kg) the response 0-No, if Yes, note the total f= Zinc (kg) Quantity of input used) g=pesticides h=herbicides i=hired labor (yes/no) j=Nutrient Expert Decision Support k=Leaf Colour Chart (yes/no) software l=GreenSeeker sensors (yes/no) i=hired labor (yes/no) 127 (choose unit mg/gm/ml/lit/kg) (choose unit mg/gm/ml/lit/kg) S.N Note to Enumerators: Start with biggest plot farmer cultivated 87 88 89 90 91 92 93 94a 94b 95 What was the Total Man Days of Labor required in the following activities for rice farming in Kharif Season 2015? Plot I Plot 2 Plot 3 land preparation Sowing Weeding Irrigating Harvesting 96-Not Applicable Other (specify) How many times was this plot irrigated in Kharif Season 2015? What is the total quantity of rice you expect to harvest from this plotif in this season? Qtl rice or dehusked: 1=husked Specify the weight is husked If plot is inter-mixed crop, ask: What is the total value of other 2=dehusked crops you expect to harvest from this plot in this season? Rs Section V: Perception on new technologies, constraints and access to information and credit 5.1. Perception of New Agriculture Practices S.N. Questions 1. 2. 3 4 CODE When was the last time your Household adopted a NEW input or a farming practice on your farm for the first time? (YYYY) What was this new input or farming practice you most recently adopted on your farm? CODE: 1-seed/Variety, 2-Agro-chemicals, 3-New animal breed, 4-agronomic practices, 5-soil or water conservation, 6-conservation agriculture, 7Machinery/tools, 8-Storage method, 9-mono-cropping, 10-drying/processing, 99Others (Specify) What is the depth of ground water level in this area?(ft) 98 Don’t know Over the past 10 years have you experienced fall in ground water level?(2- No, 98-Don’t Know, If YES, by how much ft water level has declined) If NO, skip to 6 If YES, In your opinion, what are the reasons for the decline in ground water level? CODE A of any water conservation practices? (1- Yes, 0-No) if NO, skip to Are you aware next If yes,section name the practices? (Record up to three) A B (CODE B) Are you using any of these practices on your farm? CODE: 1-Yes 2-No If YES, which one(s)? If NO, why not? CODE C Have you ever used hybrid seeds of any crop on your farm? CODE: 1-Yes 2-No 5 6 7 C 8 9a 9b 10 Code A: 1- Indiscriminate use of water, 2. Deforestation, 3- Increase in population, 4-Increased industrial activity, 5- Water Pollution, 6-Decline in rainfall, 7-Reason not related to human activity,8- Increase in submersible pump, 98-Don’t Know, 99Others (specify) Code B: 1-Scheduling irrigation only when required, 2. Planting less water requiring crops/variety, 3. Keeping residue for water conservation, 4- Adopting water saving technology, 5- Farm pond, 99. Others (specify) Code C: 1- Beyond my control, 2- Single effort will not help, 3- Water saving technology are costly, 4- My land uses surface water, 5- No water problem in my area, 99- Others (specify) 5.2 Use of technology by farmers in social network 1. How many farmers in this village/other villages you know personally and regularly interact with them on farming related issues? Would you say you personally know: 1-more than 100 farmers? 2-75-100 farmers? 3-50-75 farmers? 4-30-50 farmers? 5-20-30 farmers? 6-less than 20 farmers? 128 2. Approximately how many farmers YOU KNOW have used/currently using the following technologies on their farm? (read each, and note the numbers) Code: 997--I am not aware of this practice myself 999—I am aware about this practice but don’t know how many are using it a=Zero Tillage i Soil bunds b1=Land levelling j. Field/boundary bunds b2=Laser land levelling k. Broad bed and furrow c=DSR l. Contour bunds d=Residue retention/mulching m. Polythene mulching e=legume rotation n. Nala plugs/RFDs f=drip irrigation o. Sunken pits g=green seeker p. Farm pond h=leaf color chart q. Masonry check dams i=Nutrient expert decision support software r. Well recharge pits j=agroforestry s. Penning Sheep/Goat/Cattle h=hybrid seeds 5.3a Constraints to Wheat production S.No 1a 1b 2 3 4 5 6a 6b Questions What are the two main constraints you face in wheat First farming? CODE A Second Have you stopped planting any wheat varieties in the past 3 years that you used to grow before?(1-Yes,0-No) If NO, skip to 5.4 If yes, How many? Name the most recent variety you have discontinued What is the main reason for discontinuation? CODE B What is the current price of wheat if you were to sell it? Rs/kg What is the current price of wheat if you were to buy it? Rs/kg 5.3b Constraints to Rice production S.No 1a 1b 2 3 4 5 6a 6b 6c Questions What are the two main constraints you face in rice First farming? CODE A Second Have you stopped planting any rice varieties in the past 3 years that you used to grow before?(1-Yes,0-No) If NO, skip to 5.5 If yes, How many? Name the most recent variety you have discontinued What is the main reason for discontinuation? CODE B What is the current price of rice/paddy if you were to sell it? Rs/kg What is the current price of rice/paddy if you were to buy it? Rs/kg What is the selling price of rice you expect after harvest this season? Rs/kg Code A: 1-Land, 2-labour, 3-cash constraint, 4-seeds not available, 5-insect and disease problem, 6-cannot sell the crop, 7-price too low, 8-no information or technical advice on farming practices, 99-other(specify) Code B: 1-Seed not available, 2-had low yield, 3-did not like the color, 4-susceptible to disease, 5- not liked by processors, 6-unpleasing cooking quality/taste, 99-other (specify) 5.3d Sources of risk in farming and coping strategies 1a 1b 1c 1d Based on your experience, which of the following events would you consider to be a major cause of concern to you as a farmer or a major source of risk in your farming operation? 129 Variability in the timing and level of rainfall Floods Drought High temperatures 1e 1f 1g 1h 1i 1j 2a 2b 2c 2d 2e 2f 3 4 Read each event and ask the respondent to rank them on a scale of 0 to 2 0 – not a cause of concern for me 3- Somewhat a concern for me 4- A major concern for me Hail or cold temperatures Insects and plant diseases Infectious livestock diseases Price fluctuations in farm commodities of inputs in a Non-availability Lack market to sell the products timelyofmanner When you face an Sell household goods, jewelry, etc. economic shock due to Sell animals/ livestock any of these risk factors Sell other farm assets mentioned above, what Borrow money coping strategies do you I change my farming practices by going back to doing things the most often use? traditional I change myway practices by using NEW and MODERN methods of (indicate 1=Use most farming often, 2=use sometime, 3- ask: Can you give an example of this strategy you have used in the If use 2e often or sometime, Never) past? If use 2f often or sometime, ask: Can you give an example of this strategy you have used in the past? 5.4 Loss Due to Unexpected Weather 1. In the past 5 years, have you ever lost a significant portion of your crop production due to unexpected weather (e.g., low rainfall, flooding, unexpected monsoon time, hail, etc)? 1-Yes, 2-No 2. If yes, how many times you have suffered such losses in past five years? 98Don’t know/can’t say/don’t remember 3. How many times you have suffered such losses in the past two years? 98-Don’t know/can’t say/don’t remember 4 Which crops were most impacted by these losses? 1-wheat, 2-rice, 3groundnut, 4-other, 5-all 5a. Have you heard about the phenomenon called ‘climate change’? 1-Yes 2-No 5b If YES, can you tell me what will happen as a result of ‘climate change’ that 5c has implications for farmers like you? (select as many as mentioned): 5d 1-extreme weather; 2-too much rain/flood; 3-too little rain/drought; 4-high 5e temperatures; 5-late start of rainy season; 6-cold winters; 7-too much pest/diseases; 8-unpredictable weather; 9-farming will become more risky; 10Other (specify); 98-Don’t know 5.5 Access to information, infrastructure and credit What is the distance from your house to the nearest paved road (if the house is next 1 2 3 4 5 6a 6b 7 8 9 10 to the paved road, write zero) km What is the distance from your house to the nearest market where you obtain agricultural inputs (e.g., fertilizer, pesticides, seeds, etc.) km What is the distance from your house to the nearest agricultural extension office km What means of transport do you mainly use to get to the nearest commercial town? 1- Walking, 2- Tractor, 3- Bicycle, 4-Motorcycle, 5-Car, 6- Bus, 7- Light transport Vehicle, 8-Animal Driven Cart, 99- others (specify) Distance from your home to this commercial town km Time it takes on average to travel to this commercial town using the main Hours mode of transportation (Consider time of one way travel) minutes Do you have a bank account? Code: 1-Yes 2-No Do you own kisan credit card? (1-Yes, 2-No) Do you currently have crop insurance policy (other than KCC)? Code: 1-Yes 2-No, If YES, skip to 12 Did you have crop insurance in the past but have discontinued? Code: 1-Yes 2-No 130 11 12 13 14 15a 15b 15c 15d 16a 16b 17a 17b 18a 18b 19 20 21 22 23 Reason for not having crop insurance or for discontinuing: 1=Do not need, 2=Not available in my place, 3= No claim available at time of damage 4=Too expensive, 99= Others (Specify) Did you or anyone in the household access credit for agricultural production in the past 12 months (1=Yes, 2= No) If NO, skip to 14 If yes, from where? 1- Bank, 2- Cooperatives, 3- SHG, 4-Community member, 5-Relative/ Friend/ Neighbour, 6-Local money Lender, 7-Commission Agent, 8- Employer, 9-Agrovet, 10Trader 99- Others (Specify) If no, why not? 1= Did not need, 2= Do not have access to credit, 3=Very high interest rate , 4-Far From Residence, 5- Bank staff not cooperative, 6- was getting less amount 7-loan was not approved 8-no collateral 99= Others (Specify) Do you currently have any debt (i.e., do you owe money to anyone)? 1-Yes 2No If NO, skip to 17a If yes, to whom do you owe money? 1- Bank, 2- Cooperatives, 3- SHG, 4-Community member, 5-Relative/ Friend/ Neighbour, 6-Local money Lender, 7-Commission Agent, 8- Employer, 9-Agrovet, 10Trader 96-multiple (specify) 99- Others (Specify) What is the interest PER ANNUM you are paying on this debt? In how many years do you expect to pay-off this debt? Suppose you need to borrow money for any purpose, how likely is it that you will be able to borrow money you need? (Read the possible responses and select one) 1-Extremely likely (about 100% chance), 2-Quite likely (about 75% chance), 3-Neither likely nor unlikely (about 50%), 4-Quite unlikely (about 25% chance), 5-Extremely unlikely (about 0%) (skip to 17a) Who will be the main source of credit? 1- Bank, 2- Cooperatives, 3- SHG, 4Community member, 5-Relative/ Friend/ Neighbour, 6-Local money Lender, 7Commission Agent, 8- Employer, 9-Agrovet, 10-Trader 99- Others (Specify) During the past year, where did you receive most of your information and advice about wheat production and marketing from? CODE A (Multiple Response) In total, approximately how many times did you receive information about wheat production and marketing last year from all these sources? During the past year, where did you receive most of your information and advice about Rice production and marketing from? CODE A (Multiple Response) In total, approximately how many times did you receive information about Rice last year from all these sources? Do you use mobile phone to access information related to farming? 1=Yes 2=no If NO, go to 21 If yes, what type of information? CODE C – Multiple Response From whom do you access information using mobile phone? CODE D— Multiple Response Have you heard about the following organizations / programs? (1-Yes 2-No ) a. CIMMYT b. ICRISAT c. IRRI d. CGIAR e. CCAFS f. Climate Smart Villages g. Krishi Vignan Kendra h. Internet What is the farthest you have ever travelled? 0-never left this village; 1-a village/town in this district; 2-a village/town in a neighboring district; 3-a neighbouring state; 4-another state within India; 5-another country in South Asia; 6-US/Canada/Australia/Europe; 7-Middle east; 9Other (specify) 131 24 What is the farthest anyone who is currently a member of your household (other than you) has travelled so far and his/her relationship to you? 0-never left this village; 1-a village/town in this district; 2-a village/town a.travel b. relationship (CODE E) in a neighbouring district; 3-a neighbouring state; 4-another state within India; 5another country in South Asia; 6-US/Canada/Australia/Europe; 7-Middle east; 9Other (specify) 25 When it comes to adopting new technology, inputs or farming practices, which of the following best describes your behaviour: 1 - I am one of the first ones to adopt NEW technologies 2 – I usually wait until a few farmers I know have used those inputs/technologies/practices, and then based on their experiences I make the decision 3 – I usually wait until most farmers in this village are already using those inputs/technologies/practices, and I am 100% sure that those technologies work 4 – I rarely change my farming practices as I am not comfortable doing new things 26 Do you have life insurance policy? 1-Yes, 2-No CODE A. 1-Extension agent, 2-NGO staff, 3-Trader / input dealer, 4-Farmer group/leader farmer, 5-Service Provider, 6-I did not receive any information or advice, 99-other (specify) CODE B. 1=1-2 times, 2=2-3 times, 3=3-5 times, 4=5-10 times, 5=More than 10 times CODE C. 1-Weather, 2-Price, 3-Inputs, 4-Production technology, 5-Pest control, 6-government programs, 99other (Specify) CODE D. 1-Relatives/friends, 2-Input dealers, 3-KVK, 4-Kisan Call Center, 5-Extension agents, 6-RML, 7-IFFCO Kisan Sanchar Limited (IKSL), 8-mKRISHI, 9-other mobile based agro advisory services, 99-Other (specify) Code A: 2-wife, 3-Husband, 4- Son, 5-Daughter, 6- Grandchild, 7- Father, 8-Mother, 9- Sister, 10-Brother, 11- Niece, 12Nephew, 13- Son in law, 14- Daughter in law, 15-Brother in law, 16-Sister in Law, 17- Father in law, 18- Mother in law 132 5.6 Assets owned How many of the following does your household own and what is the total value (in Rs) if you were to sell it today? (Instruction: For each item, write the number owned and its total value across all units owned. If none owned, write zero) 1 2 3 4 5 6 7 9 10 Bicycle Motor cycle / Scooter Car / truck Cart Tractor Plough Metal silos Water tank 8 Irrigation / water pump Greenhou Dehusker se / glass house 11 12 Fodder chopper 13 14 15 16 17 18 19 20 21 22 23 24 Combine harvester Cultivator / tiller Zero till drills Biogas plant Turbo/ Happy seeder Seed-cum- LLL Ferti Drills Ripper Radio / cassette player TV Fans AC 25 26 27 28 29 30 31 32 33 34 35 36 Cooler Washing machine Water purifier Camera Pressure cooker Almirah Refrigerator Comp-uter Sewing machine Gas stove Mobile phones NonMobile phone # owned value # owned value # owned value How many animals does your household currently own and its total value across all units owned (if none owned, write zero) 36 38 39 40 41 42 43 44 45 Horses Cows Buffaloes Bulls Goats Sheep #value owned To be filled by Enumerator after the completion of Survey End Time Was the survey completed in first attempt or required a revisit? 1-First Attempt, 2-ReVisit 133 Donkeys / Mules Pigs Chickens 46 Other (describe)